CN106897969B - A data processing device and method for super-resolution positioning microscopic imaging - Google Patents

Abstract

The invention discloses a data processing device and method for super-resolution positioning microscopic imaging. The data processing device includes a data preprocessing circuit and a first processor. The data preprocessing circuit includes a detector interface circuit, a multiplexing circuit, FPGA and a first interface circuit. The detector interface circuit quickly acquires and outputs the fluorescence map collected by the detector. The multiplexing circuit replicates the fluorescence image collected by the detector, at least one path is a path, and the original circuit structure is not affected. The FPGA extracts the fluorescent molecule sub-regions from the fluorescence image, and the first interface circuit transmits the fluorescent molecule sub-regions to the first processor. The first processor performs positioning processing on the fluorescent molecule sub-regions to obtain a super-resolution reconstruction image. Since the FPGA can quickly extract fluorescent molecular subregions from the fluorescence image and transmit them to the first processor, the super-resolution reconstructed image is obtained by the first processor, so that the data processing device realizes real-time reconstruction of high-precision super-resolution images.

Description

A data processing device and method for super-resolution positioning microscopic imaging

technical field

The invention belongs to the technical field of super-resolution positioning microscopic imaging, and more particularly relates to a data processing device and method for super-resolution positioning microscopic imaging.

Background technique

Super-resolution positioning imaging technology can achieve a spatial resolution of 20nm, and can study the complex working mechanism in cells from the molecular level. It has become an indispensable research tool in the field of life science research. With the rapid development of detector technology, people hope to improve the temporal resolution of super-resolution localization imaging without sacrificing the imaging field of view and spatial resolution. In order to improve the time resolution of super-resolution localization microscopy imaging, the rapidly developing low-light detector (sCMOS) has achieved a theoretical image acquisition rate of 420M pixel/s. However, at present, the computing speed of the collected data is far from being able to be processed in real time by the computer. The current main system architecture is to directly connect the data collected by the sCMOS detector to the computer, and the computer's central processing unit (CPU) and graphics processing unit (GPU) jointly calculate, and realize image processing through super-resolution positioning algorithms. According to different algorithms, the accuracy and speed of the obtained super-resolution images vary greatly. But none of them can realize real-time processing. For sCMOS detectors, the rate of image data acquisition far exceeds the calculation rate of the algorithm. Even for algebraic algorithms with low precision, it takes a long time for data processing to obtain a super-resolution reconstruction map, which greatly restricts the development of super-resolution positioning microscopy imaging technology. Therefore, it is an urgent need in this field to invent a method that can satisfy both high-precision spatial resolution and real-time data processing.

Contents of the invention

Aiming at the defects of the prior art, the present invention provides a data processing device for super-resolution positioning microscopic imaging, which aims to solve the technical problem that the existing processing device cannot take into account high-precision spatial resolution and satisfy real-time data processing .

To achieve the above object, the present invention provides a data processing device for super-resolution positioning microscopic imaging, including a data preprocessing circuit and a first processor;

The data preprocessing circuit includes:

Detector interface circuit for collecting and transmitting fluorescence images;

an FPGA, the input of which is connected to the output of the detector interface circuit for extracting fluorescent molecular subregions from the fluorescence map; and

The first interface circuit, whose input end is connected to the output end of the FPGA, is used to transmit the fluorescent molecule sub-region to the first processor;

The first processor includes:

A memory whose first end is connected to the output end of the first interface circuit for storing the fluorescent molecule sub-regions;

The CPU is used to send positioning instructions to the GPU and receive super-resolution reconstruction images output by the GPU;

GPU, whose first end is connected to the first end of the CPU, and whose second end is connected to the second end of the memory, is used to locate the fluorescent molecule sub-region according to the positioning instruction, obtain a super-resolution reconstruction map, and super-resolution The reconstructed image is transmitted to the CPU, and the super-resolution reconstructed image is finally displayed on the image display.

In the data processing device provided by the present invention, the data preprocessing circuit extracts the fluorescent molecule sub-regions from the fluorescent image, and transmits the fluorescent molecule sub-regions to the first processor, and the CPU and GPU in the first processor only complete the fluorescent molecular sub-regions. The region positioning processing can improve the efficiency of the first processor in obtaining the super-resolution reconstruction image. The fluorescence image can be quickly obtained from the detector through the detector interface circuit. The hardware circuit provided by the FPGA can extract the fluorescent molecular sub-region with a small amount of data from a large number of fluorescent images in real time, and the fluorescent molecular sub-region can be extracted by the first interface circuit. The region is quickly transferred to the first processor. Therefore, the data processing device provided by the present invention can realize real-time reconstruction of high-precision super-resolution images.

Further, the data preprocessing circuit also includes: a multiplexing circuit, its input terminal is connected to the output terminal of the detector interface circuit, its first output terminal is connected to the input terminal of the FPGA, and its second output terminal is used to convert the fluorescent The map is transmitted to a second processor, and multiplexing circuitry is used to receive and duplicate the fluorescence map into two fluorescence map outputs.

The multiplexing circuit realizes dividing the fluorescent image into multiple fluorescent images, one fluorescent image is transmitted to the data preprocessing circuit, and one fluorescent image is transmitted to the second processor, and the second processor performs super-resolution positioning processing on the fluorescent image, The data processing device provided by the present invention is compatible with the existing data processing device, and the user can retain the original image processing method.

Further, the data preprocessing circuit also includes: a multiplexing circuit, its input terminal is connected to the output terminal of the detector interface circuit, its first output terminal is connected to the input terminal of the FPGA, and its second output terminal is used to convert the fluorescent image It is transmitted to the second processor, and the third terminal to the Nth terminal thereof are used as expansion interfaces, and the multiplexing circuit is used to receive and copy the fluorescence image into multiple output channels of the fluorescence image, wherein, N≥3.

Further, FPGA includes:

Fluorescence image reading module, the input end of which is connected to the output end of the detector interface circuit, used to obtain the fluorescence image, and divide the fluorescence image into three channels of fluorescence image output;

A noise reduction processing module, the input end of which is connected to the first output end of the fluorescence image reading module, and is used to perform noise reduction processing on the fluorescence image and output the first image;

Remove the background processing module, its input end is connected with the output end of the noise reduction processing module, for carrying out the background processing to the first image, output the second image;

The background fluctuation intensity acquisition module includes a third sub-region acquisition circuit, the input end of which is connected to the second output end of the fluorescence image reading module, and is used to extract the third current processing sub-region from the fluorescence image; the background fluctuation intensity acquisition circuit, Its input end is connected with the output end of the third sub-region obtaining circuit, is used for obtaining the local standard deviation of the 3rd current processing sub-region, and uses the local standard deviation of the 3rd current processing sub-region as the background of the 3rd current processing sub-region Fluctuating intensity, used as a threshold for determination of fluorescent molecule sub-regions; and

The sub-region judgment and extraction module includes a fourth sub-region acquisition circuit, whose input terminal is connected to the output end of the background processing module, and is used to extract the fourth current processing sub-region from the second image; the sub-region judgment circuit, whose first One input end is connected with the output end of the fourth sub-area acquisition circuit, and its second input end is connected with the output end of the background fluctuation intensity acquisition circuit, and the background fluctuation intensity of the third current processing sub-area is used as the fourth current processing sub-area Background fluctuation intensity, according to the background fluctuation intensity of the fourth current processing sub-region and the signal strength of the fourth current processing sub-region, determine whether there are fluorescent molecules in the fourth current processing sub-region, and output the sub-region extraction control signal; alternative sub-region acquisition A circuit, whose input end is connected to the third output end of the fluorescent image reading module, and is used to extract an alternative subregion from the fluorescent image; a subregion extraction circuit, whose input end is connected to an output end of the optional subregion extraction circuit , the control end of which is connected to the output end of the sub-region judging circuit, and determines whether the candidate sub-region is a fluorescent molecule sub-region according to the sub-region extraction control signal.

In the background fluctuation intensity acquisition module, the third current processing sub-area is processed with local standard deviation to obtain the background fluctuation intensity of the third current processing sub-area, and only local image information is used, so that each sub-area has a background fluctuation intensity, Instead of all sub-regions in the entire image sharing a background fluctuation intensity, the extracted fluorescent molecule sub-regions can be more accurate and have better adaptability when the entire image has an uneven background, which is completed by the circuit in the FPGA. Compared with the existing data processing device that uses CPU and GPU to jointly process the fluorescent image, the data processing rate is significantly improved by extracting the fluorescent molecular sub-region from the fluorescent image.

Further, the background fluctuation intensity acquisition circuit includes:

A pixel comparator, whose input end is connected to the output end of the third sub-area acquisition circuit, and is used to obtain screened pixels from the third current processing sub-area according to the gray value of each pixel in the third current processing sub-area;

A mean value adder, whose input is connected to the output of the pixel comparator, for obtaining the mean value of the screened pixels;

a subtractor, the first input of which is connected to the output of the pixel comparator, and the second input of which is connected to the output of the mean value adder, for obtaining the difference between each screened pixel and the average value of the screened pixel; and

The local standard deviation adder, whose input terminal is connected to the output terminal of the subtractor, is used for summing the difference between each screened pixel and the average value of the screened pixel to obtain the local standard deviation of the third currently processed sub-region.

The background fluctuation intensity acquisition circuit provided by the present invention uses the pixel comparator to obtain screened pixels according to the gray value of the third current processing sub-region, and uses the screened pixels to calculate the local standard deviation, which can improve the processing speed of the background fluctuation intensity acquisition circuit.

Further, the subregion judging circuit includes:

The first pixel comparator, whose input terminal is connected to the output terminal of the fourth sub-region acquisition circuit, is used to judge whether the central pixel gray value of the fourth current processing sub-region is the gray value of each pixel in the fourth current processing sub-region value, and output the first level value according to the judgment result;

A second pixel comparator, one input end of which is connected to the output end of the fourth sub-area acquisition circuit, and the other input end is connected to the output end of the background fluctuation intensity acquisition circuit, for processing according to the fourth current processing sub-area The central pixel gray value and the background fluctuation intensity of the fourth currently processed sub-region output a second level value;

The four-neighborhood pixel adder, whose input end is connected to the output end of the fourth sub-area acquisition circuit, is used to combine the four-neighborhood pixel gray value of the central pixel of the fourth current processing sub-area and the central pixel of the fourth current processing sub-area The gray value is accumulated and processed to output the first accumulated gray value;

An eight-neighborhood pixel adder, whose input end is connected to the output end of the fourth sub-area acquisition circuit, is used to combine the eight-neighborhood pixel gray value of the center pixel of the fourth current processing sub-area and the gray value of the fourth current processing sub-area The central pixel gray value is accumulated and processed to obtain a second accumulated gray value;

A four-neighborhood pixel comparator, one input end of which is connected to the output end of the four-neighborhood pixel adder, and the other input end thereof is connected to the output end of the background fluctuation intensity acquisition circuit, and is used for accumulating gray values according to the first sum and The background fluctuation intensity of the fourth currently processed sub-region outputs a third level value;

An eight-neighborhood pixel comparator, one input end of which is connected to the output end of the eight-neighborhood pixel adder, and the other input end is connected to the output end of the background fluctuation intensity acquisition circuit, which is used to obtain and the background fluctuation intensity of the fourth currently processed sub-region outputs a fourth level value; and

A logic AND gate, its first input end is connected to the output end of the first pixel comparator, its second input end is connected to the output end of the second pixel comparator, and its third input end is connected to the four neighbors The output terminal of the pixel comparator is connected, and the fourth input terminal thereof is connected to the output terminal of the eight-neighborhood pixel comparator for outputting sub-region extraction control signals according to the first level value to the fourth level value.

The sub-region judging and extracting module provided by the present invention determines whether the fluorescent molecule is located in the fourth current processing sub-region by judging whether the gray value of the central pixel of the fourth current processing sub-region is the maximum value of all gray values of the fourth current processing sub-region. For the central position of the sub-region, the background fluctuation intensity of the fourth current processing sub-region is used as the threshold value, and the gray value of the central pixel of the fourth current processing sub-region, the gray value of the four neighboring pixels, the gray value of the eight neighboring pixels The degree value is compared with the background fluctuation intensity of the fourth current processing sub-region to determine whether there are fluorescent molecules in the fourth current processing sub-region, and the influence of the background on the determination of fluorescent molecules is eliminated, so that the judgment result is more accurate.

As another aspect of the present invention, the present invention provides a data preprocessing method for super-resolution positioning microscopic imaging, comprising the following steps:

(1) performing noise reduction processing on the fluorescence image to output the first image, and performing background noise removal processing on the first image, and outputting the second image;

Extract the third current processing sub-region from the fluorescence map, and perform local standard deviation processing on the third current processing sub-region, obtain the local standard deviation of the third current processing sub-region, and convert the local standard deviation of the third current processing sub-region As the background fluctuation intensity of the third current processing sub-area; update the third current processing sub-area, obtain the background fluctuation intensity of each pixel in the fluorescence map, and use the background fluctuation intensity of all pixels in the fluorescence map as the background fluctuation of all pixels in the second image strength;

(2) Extract the fourth current processing sub-region from the second image, use the background fluctuation intensity of the third current processing sub-region as the background fluctuation intensity of the fourth current processing sub-region, according to the signal strength of the fourth current processing sub-region and the first Determine whether there is a fluorescent molecule sub-region in the fourth current processing sub-region according to the background fluctuation intensity of the fourth current processing sub-region, and determine whether to use the same region in the fluorescence image as the fourth current processing sub-region in the second image according to the judgment result. Fluorescent molecule sub-region; update the fourth current processing sub-region, and extract all fluorescent molecule sub-regions in the fluorescence map;

(3) Perform positioning processing on the fluorescent molecular sub-regions to obtain super-resolution reconstructed images.

In the data preprocessing method for super-resolution positioning microscopic imaging provided by the present invention, each third current processing sub-region extracted from the fluorescence map is subjected to local standard deviation processing to obtain the background fluctuation intensity of the third current processing sub-region , and extract the fourth current processing sub-region from the second image, use the background fluctuation intensity of the third current processing sub-region as the background fluctuation intensity of the fourth current processing sub-region, and according to the signal strength of the fourth current processing sub-region and The background fluctuation intensity of the fourth current processing sub-region determines whether there are fluorescent molecules in the fourth current processing sub-region, which can make the extracted fluorescent molecule sub-regions more accurate and have better adaptability when the entire image has an uneven background. to obtain accurate super-resolution reconstruction images.

Further, step (2) includes the following steps:

(21) extract the fourth current processing sub-region from the second image, if the gray value of the central pixel of the fourth current processing sub-region is the maximum value of the gray value of all pixels in the fourth current processing sub-region,

And the gray value of the central pixel of the fourth current processing sub-region is greater than twice the background fluctuation intensity of the fourth current processing sub-region;

And the sum of the gray value of the central pixel of the fourth current processing sub-region and the gray value of the four neighboring pixels of the central pixel of the fourth current processing sub-region is greater than 9 times of the background fluctuation intensity of the fourth current processing sub-region,

And the sum of the gray value of the central pixel of the fourth current processing sub-region and the gray value of the eight neighboring pixels of the central pixel of the fourth current processing sub-region is greater than 11 times of the background fluctuation intensity of the fourth current processing sub-region;

Then extract the region with the same position in the second image as the fluorescent molecule sub-region in the fluorescence image and the fourth current processing sub-region, and enter step (22); otherwise do not extract the fourth current processing sub-region in the fluorescence image and the fourth current processing sub-region For the same region in the two images, enter step (22);

(22) Judging whether all the 4th current processing sub-regions have been extracted, if so, then terminate; otherwise, enter step (21);

The fourth currently processed sub-area is a sub-area larger than 7×7 extracted from the second image.

Extract the fourth current processing sub-region from the second image, compare the gray value of each pixel in the fourth current processing sub-region with the background fluctuation intensity of the fourth current processing sub-region, and determine whether the fourth current processing sub-region is fluorescent molecule, by judging whether the gray value of the center pixel of the fourth current processing sub-region is the maximum value to determine whether the fluorescent molecule is in the center of the fourth current processing sub-region, the area where the fluorescent molecule is located in the fluorescence map can be accurately judged.

Description of drawings

Fig. 1 is a schematic structural view of a data processing device for super-resolution positioning microscopic imaging provided by the present invention;

2 is a schematic diagram of an FPGA preprocessing circuit in a data processing device for super-resolution positioning microscopic imaging provided by the present invention;

Fig. 3 is the schematic structural diagram of the background fluctuation intensity acquisition circuit in the FPGA;

Fig. 4 is that the present invention provides the structural representation of FPGA neutron region judgment circuit;

Fig. 5 is the structure schematic diagram of the FPGA device provided by the present invention, (a) is the structure diagram that obtains the fluorescence figure from sCMOS detector, (b) is the structure diagram of the image preprocessor in FPGA;

Fig. 6 is a schematic diagram of obtaining screening pixels from the third current processing sub-region in the data processing method provided by the present invention;

Fig. 7 is a schematic diagram of the effect of the fluorescence image preprocessing provided by the data preprocessing circuit provided by the present invention, wherein (a) is a schematic diagram of the effect of the fluorescence image collected by the sCOMS camera, and (b) is the noise reduction processing and background removal of the fluorescence image Schematic diagram of the effect of processing, (c) is a schematic diagram of the effect of background fluctuation intensity after local standard deviation filtering of the fluorescence image, (d) is a schematic diagram of the effect of all fluorescent molecular sub-regions extracted from the fluorescence image.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

Fig. 1 is the structural representation of the data processing device that is used for super-resolution positioning microscopic imaging that the present invention provides, and data processing device comprises data preprocessing circuit, and the output end of data preprocessing circuit is connected the input end of the first processor, and data preprocessing circuit The processing circuit is used to extract the fluorescent molecule sub-regions from the fluorescence image collected by the detector, and the first processor is used to perform positioning processing on the fluorescent molecule sub-regions to obtain a super-resolution reconstruction image. The data preprocessing circuit includes a detector interface circuit, which is used to quickly transmit the fluorescence image collected by the detector to the multiplexing circuit. The detector interface circuit can be various transferable interfaces such as cameralink interface, USB interface, and FMC module.

The multiplexing circuit is used to receive the fluorescent image and copy the fluorescent image into a multi-channel fluorescent image output, and its input terminal is connected to the output terminal of the detector interface circuit. The multiplexing circuit has two output terminals, one of which is The path is used to connect with the input end of the second processor, the other output end of the multiplexing circuit is connected with the FPGA input end, the output end of the second processor is connected with the input end of the image display, and the second processing The device is used to perform super-resolution positioning processing on the fluorescence image to obtain a super-resolution image, and the multiplexing circuit allows the user to retain the original processing method of the fluorescence image.

As another implementation of the multiplexing circuit, the multiplexing circuit has N output terminals, wherein N≥3, the first output terminal is connected to the input terminal of the FPGA, and the second output terminal is connected to the second processor's The input terminal is connected, and the third output terminal to the Nth output terminal are all used for expanding the output. Other processing chips or processors are connected to the data preprocessing circuit through the third output terminal to the Nth output terminal, and the processing chip may be a DSP chip.

The FPGA output terminal and the input terminal of the first interface circuit, the FPGA extracts the fluorescent molecule sub-region from the fluorescence map, and the fluorescent molecule sub-region is transmitted to the first processor by the first interface circuit, and the first processor includes a memory, a CPU and a GPU, The first end of the memory is connected to the output end of the first interface circuit, the first end of the CPU is connected to the first end of the GPU, the second end of the GPU is connected to the second end of the memory, the memory is used to store fluorescent molecule sub-regions, and the CPU uses After sending the positioning instruction to the GPU, the GPU receives the fluorescent molecule sub-region data from the memory after receiving the positioning instruction, and performs positioning processing on the fluorescent molecule sub-region to obtain a super-resolution reconstruction image, and transmits the super-resolution image to the CPU, and displays it on the image display Show the super-resolution reconstruction map.

Due to the sparse excitation characteristics of super-resolution positioning imaging, only the sub-regions containing fluorescent molecules in the fluorescence image collected by the receiving detector are effective. Therefore, the data processing device provided by the present invention is extracted from the fluorescence image by the data preprocessing circuit. Fluorescent molecular sub-regions, the fluorescent molecular sub-regions are transmitted to the first processor, and the CPU and GPU in the first processor only complete the positioning processing of the fluorescent molecular sub-regions, which can improve the efficiency of the first processor in obtaining super-resolution reconstruction images . The fluorescence image is quickly obtained from the detector through the detector interface circuit, and the fluorescent molecule sub-regions in the fluorescence image are extracted in real time by FPGA, and the useful fluorescent molecule sub-regions in the fluorescence image are proposed, which can ensure the accuracy of super-resolution reconstruction images , the hardware circuit provided inside the FPGA can extract fluorescent molecule subregions from a large number of fluorescent images in real time, and the fluorescent molecule subregions are quickly transmitted to the first processor by the first interface circuit. Therefore, the data processing device for super-resolution positioning microscopic imaging provided by the present invention realizes real-time reconstruction of high-precision super-resolution images.

Figure 2 is a schematic diagram of the FPGA structure in the data processing device for super-resolution positioning microscopic imaging provided by the present invention, the FPGA includes a fluorescence image reading module 5, a noise reduction processing module 1, a background processing module 2, and a background fluctuation intensity acquisition module 3 and the sub-region judgment and extraction module 4.

The fluorescence image reading module 5 is used to obtain the fluorescence image, and divide the fluorescence image into three-way fluorescence image output, the first output end of the fluorescence image reading module 5 is connected to the input end of the noise reduction processing module 1, and the noise reduction processing Module 1 is used to perform noise reduction processing on the fluorescence image, and output the first image. The output terminal of the noise reduction processing module 1 is connected to the input terminal of the background removal processing module 2, and is used to perform background removal processing on the first image, and output the second image. image.

The background fluctuation intensity acquisition module 3 includes a third sub-region acquisition circuit 301 and a background fluctuation intensity acquisition circuit 302, and the input end of the third sub-region acquisition circuit 301 is connected with the output end of the fluorescence image reading module 5 for real-time acquisition from the fluorescence The third current processing sub-region is extracted in the figure, the input end of the background fluctuation intensity acquisition circuit 302 is connected with the output end of the third sub-region acquisition circuit 301, for real-time output of the local standard deviation of the third current processing sub-region, and The local standard deviation of the third current processing sub-region is used as the background fluctuation intensity of the third current processing sub-region, and the background fluctuation intensity is used as a threshold for judging the fluorescent molecule sub-region.

The sub-region judgment and extraction module 4 includes a fourth sub-region acquisition circuit 401, a sub-region judgment circuit 402, an alternative sub-region acquisition circuit 403 and a sub-region extraction circuit 404, the input terminal of the fourth sub-region acquisition circuit 401 is connected to the background The output end of the processing module 2 is connected to extract the fourth current processing sub-region from the second image in real time, the first input end of the sub-region judging circuit 402 is connected to the output end of the fourth sub-region obtaining circuit 401, and the sub-region judging circuit 402 The second input end is connected to the output end of the background fluctuation intensity acquisition circuit 302, so that the fourth sub-area acquisition circuit 401 and the third sub-area acquisition circuit 301 can cooperate in timing, so that the central pixel of the fourth current processing sub-area The position in the second image is the same as that of the center pixel of the third current processing sub-region in the fluorescence map, that is, the background fluctuation intensity of the third current processing sub-region is used as the background fluctuation intensity of the fourth current processing sub-region, and the sub-region judgment circuit 402 according to The fourth current processing sub-region and the background fluctuation intensity of the third current processing sub-region output by the background fluctuation intensity acquisition circuit 302 judge in real time whether there are fluorescent molecules in the fourth current processing sub-region, and output a sub-region extraction control signal.

The input terminal of the candidate subregion acquisition circuit 403 is connected with the third output terminal of the fluorescence image reading module 5, the alternative subregion acquisition circuit 403 is used to extract the candidate subregion from the fluorescence image, and the input terminal of the subregion extraction circuit 404 It is connected to the output terminal of the candidate subregion acquisition circuit 403, the control terminal of the subregion extraction circuit 404 is connected to the output terminal of the subregion judgment circuit 402, and determines whether the candidate subregion is a fluorescent molecule subregion in real time according to the subregion extraction control signal. Allow the candidate subregion extraction circuit 403 to cooperate with the fourth subregion acquisition circuit 401 in timing, so that the position of the central pixel of the fourth currently processed subregion in the second image is the same as the position of the central pixel of the candidate subregion in the fluorescence map same.

Through the third sub-region acquisition circuit 301, the fourth sub-region acquisition circuit 401 and the alternative sub-region extraction circuit 403, the third current processing sub-region, the fourth current processing sub-region and the alternative sub-region are continuously updated to realize the fluorescence detection. All pixels in the image are preprocessed, and the fluorescence image is processed in real time through the data preprocessing circuit to reduce the data in the fluorescence image that has no meaning or bad influence on the reconstructed image, and the fluorescent molecular sub-region is used for image reconstruction, effectively reducing the calculation of image reconstruction The amount of calculation reduces the time to reconstruct the image.

At the same time, the background fluctuation intensity acquisition module obtains the local standard deviation of the third current processing sub-region according to the gray value of the third current processing sub-region, only uses the local image information, and uses the local standard deviation of the third current processing sub-region As the background fluctuation intensity of the third current processing sub-region, each pixel has a value of background fluctuation intensity, rather than the whole image sharing a value of background fluctuation intensity, which can make the whole image have an uneven background The extracted fluorescent molecular subregions are more accurate and have better adaptability. Therefore, the data processing device for super-resolution positioning microscopic imaging provided by the present invention can realize real-time and high-precision spatial resolution reconstruction of images.

The noise reduction processing module in the FPGA provided by the present invention includes a noise reduction row buffer group, a first noise reduction register group, a noise reduction multiplication and accumulation circuit and a second noise reduction register, and the noise reduction row buffer group is used to extract the noise from the fluorescent image. Extract and output the first current processing sub-area, the input end of the first noise reduction register group is connected with the output end of the noise reduction line buffer group, the first noise reduction register group is used to store the first current processing sub-area, and the noise reduction is multiplied The input end of the accumulation circuit is connected to the output end of the first noise reduction register group, and the noise reduction multiplication and accumulation circuit is used to perform multiplication and accumulation processing on the first current processing sub-region and the preset noise reduction convolution template, and output the first current To process the noise-reduced gray value of the sub-region, the input end of the second noise reduction register is connected to the output end of the noise reduction multiplication and accumulation circuit for storing the noise-reduced image output by the noise reduction multiplication and accumulation circuit.

In the FPGA provided by the present invention, the background processing module includes removing the background row buffer group, the first background register group, the background multiplication and accumulation circuit and the second background register, and the background row buffer group is used for noise reduction Extract and output the first current processing sub-area from the image after, the input end of the first background removing register group is connected with the output end of the background line buffer group, the first background removing register group is used to store the first current processing sub-region , the input terminal of the background removal multiplication and accumulation circuit is connected to the output end of the first background removal register group, and the background removal multiplication and accumulation circuit is used to perform multiplication and accumulation processing on the first current processing sub-region and the preset background removal convolution template, And output the gray value after removing the background of the first current processing sub-region, the input end of the second background register is connected with the output end of the background multiplying and accumulating circuit, and is used for storing the background removing outputted by the background multiplying and accumulating circuit And the image data after noise reduction.

As shown in Figure 3, the background fluctuation intensity acquisition circuit 302 in the FPGA provided by the present invention includes a pixel comparator 3021, an average value adder 3022, a subtractor 3023 and a local standard deviation adder 3024, and the input terminal of the pixel comparator 3021 is connected to the third sub- The output terminal of the area acquisition circuit 301 is connected to obtain the screened pixel according to the gray value of each pixel in the third current processing sub-area, and the input terminal of the mean value adder 3022 is connected to the output terminal of the pixel comparator 3021 to obtain the screened pixel. Average value, the first input end of the subtractor 3023 is connected to the output end of the pixel comparator 3021, the second input end of the subtractor 3023 is connected to the output end of the mean value adder 3022, and is used to obtain the average value of each screening pixel and the screening pixel Difference, local standard deviation The input terminal of the adder 3024 is connected to the output terminal of the subtractor 3023, and is used for summing the difference between each screened pixel and the average value of the screened pixel to obtain the local standard deviation of the third current processing sub-region. When the pixel comparator 3021 obtains the filtered pixels from the third current processing sub-region, it can divide all the pixels in the outermost layer of the third current processing sub-region into two parts according to the position of the pixels, and select half of the pixels from each part As screening pixels, the gray values of the pixels screened out from the part are all smaller than the gray values of any pixels not screened out in the part. The background fluctuation intensity acquisition circuit provided by the present invention uses the pixel comparator to obtain screened pixels according to the gray value of the third current processing sub-area, and uses the screened pixels to calculate the local standard deviation, which can improve the processing speed of the circuit.

As shown in Figure 4, the sub-region judgment and extraction module in the FPGA provided by the present invention includes a first pixel comparator 4021, a second pixel comparator 4022, a four-neighborhood pixel adder 4023, an eight-neighborhood pixel adder 4025, and a four-neighborhood pixel adder 4025. Comparator 4024, eight-neighborhood pixel comparator 4026 and logical AND gate 4027, the input end of the first pixel comparator 4021 is connected to the output end of the fourth sub-area acquisition circuit 401, and is used to determine the center pixel of the fourth current processing sub-area Whether the gray value is the maximum value among the gray values of each pixel in the fourth current sub-area, and output the first level according to the judgment result, when the gray value of the center pixel of the fourth current processing sub-area is the maximum value of each pixel in the fourth current sub-area If the gray value is the maximum value, the first level is a high level, otherwise, it is a low level, and whether the fluorescent molecule is located in the center of the fourth current processing sub-region can be determined through the first level.

One input end of the second pixel comparator 4022 is connected to the output end of the fourth sub-region acquisition circuit 401, the other input end of the second pixel comparator 4022 is connected to the output end of the background fluctuation intensity acquisition circuit 302, by comparing the fourth current processing The central pixel gray value of the sub-region and the background fluctuation intensity of the fourth current processing sub-region output the second level, if the central pixel gray value of the fourth current processing sub-region is greater than the background fluctuation intensity of the fourth current processing sub-region twice, the second level is high level, otherwise, the second level is low level.

The input end of the four-neighborhood pixel adder 4023 is connected to the output end of the fourth sub-region acquisition circuit 401, and is used to combine the four-neighborhood pixel gray value of the central pixel of the fourth current processing sub-region and the central pixel gray value of the fourth current processing sub-region grayscale values to obtain the first accumulated grayscale value. The input end of the eight-neighborhood pixel adder 4025 is connected to the output end of the fourth sub-area acquisition circuit 401, and is used to convert the eight-neighborhood pixel gray value of the central pixel of the fourth current processing sub-area to the center of the fourth current processing sub-area. The pixel grayscale values are accumulated to obtain a second accumulated grayscale value.

One input end of the four-neighborhood pixel comparator 4024 is connected to the output end of the four-neighborhood pixel adder 4023, and the other input end of the four-neighborhood pixel comparator 4024 is connected to the output end of the background fluctuation intensity acquisition circuit 302 to compare the first accumulated gray value and The background fluctuation intensity of the fourth current processing sub-region, and output the third level according to the comparison result, if the first accumulative element gray value is greater than 9 times of the background fluctuation intensity of the fourth current processing sub-region, then the third level is High level, otherwise, the third level is low level.

One input end of the eight-neighborhood pixel comparator 4026 is connected to the output end of the eight-neighborhood pixel adder 4025, and the other input end of the eight-neighborhood pixel comparator 4026 is connected to the output end of the background fluctuation intensity acquisition circuit 302 to compare the second accumulated gray intensity value and the background fluctuation intensity of the fourth current processing sub-region, and output the fourth level according to the comparison result, if the second accumulated element gray value is greater than 11 times of the background fluctuation intensity of the fourth current processing sub-region, then the fourth level is high level, otherwise, the fourth level is low level.

The first input terminal of the logical AND gate 4027 is connected to the output terminal of the first pixel comparator 4021, the second input terminal of the logical AND gate 4027 is connected to the output terminal of the second pixel comparator 4022, and the third input terminal of the logical AND gate 4027 is connected to the four adjacent The output terminal of the domain pixel comparator 4024 is connected, the fourth input terminal of the logical AND gate 4027 is connected with the output terminal of the eight-neighborhood pixel comparator 4026, and the sub-region extraction control signal is output according to the first level to the fourth level.

When the first level to the fourth level are all high levels, it means that there are fluorescent molecules in the fourth current processing sub-region, and the sub-region control signal output by the logic AND gate 4027 makes the alternative sub-region obtain the alternative sub-region output by the circuit. Regions are output as fluorescent molecule subregions via a subregion extraction circuit. When one of the first level to the fourth level is not high, it means that there is no fluorescent molecule in the fourth current processing sub-area, and the sub-area control signal output by the logical AND gate 4027 makes the alternative sub-area acquisition circuit output The alternative subregions of are not output via the subregion extraction circuit. By comparing the four-neighborhood pixels and eight-neighborhood pixels of the fourth current processing sub-area with the background fluctuation intensity of the fourth current processing sub-area, the influence of the background fluctuation intensity on the judgment of fluorescent molecules can be eliminated, so that the output fluorescent molecule sub-area more precise.

In the embodiment of the data processing device provided by the present invention, the detector is an sCMOS detector, and the sCMOS detector uses an 80-bits camera link transmission protocol to transmit the fluorescence image, that is, transmits data of 5 pixels at the same time per clock. Therefore, the cameralink interface circuit is selected for the interface. Figure 5 is a schematic diagram of the device structure of the PFGA. In order to correctly obtain the image data transmitted by the cameralink interface, the circuit structure shown in 5(a) is used to obtain the fluorescence image from the sCMOS detector. The sCMOS detector usually adopts the transmission mode of Rolling shutter, that is, the fluorescence image is read out line by line from the middle to both sides, and the fluorescence image is stored in the odd-line image FIFO and the even-line image FIFO by writing the state machine, while considering the fluorescence image processing It is often impossible to process a few pixels on the edge. If only a simple parity segmentation is performed, there will be a black band in the middle of the final super-resolution image. Therefore, we also transmit the 6 rows of data in the middle of the upper half of the detectors to the lower half of the detectors, and discard the 6 rows of data at the bottom of the lower half of the detectors. A similar operation is performed on the data of the upper half of the detectors, so that not only the black band in the middle is removed, but also the speed of image processing is guaranteed because the image size remains unchanged, and the fluorescence image is divided according to the parity of the number of lines in the fluorescence image. Divide the data into the upper and lower halves of the detectors.

As shown in Figure 5(b), each half-detector data is assigned an image processor for processing, and each image processor includes a noise reduction processing module, a background processing module, a background fluctuation intensity acquisition module, and sub-region judgment and extraction Module, the above four modules all include interconnected row buffer groups and register groups. In a row buffer, each clock pixel moves forward by one frame, and the pixels that are moved out of the current row buffer flow into the next row at the same time. Buffers and registers, the register group is filled with the gray value of the incoming pixels, that is, the current processing sub-region is obtained, and the current processing sub-region extracted from the fluorescent image by the line buffer group and the register group in the noise reduction processing module is consistent with the preset reduction The noise convolution template is multiplied and accumulated to obtain the gray value of the current processing pixel after noise reduction, that is, by performing Gaussian low-pass filtering on the current processing sub-region, the gray value of the current processing pixel after noise reduction is obtained, and each pixel The gray value after noise reduction is stored by the register group, and the register group outputs the gray value after noise reduction of each pixel to the background processing module. Similarly, the line buffer group and register group extract the current processed image from the first image. Sub-area, the current processing sub-area output by the register group in the background removal processing module is multiplied and accumulated with the preset background removal convolution template to obtain the gray value of the current processing pixel after background removal and noise reduction, that is, the current processing pixel Perform ring filter processing to obtain the gray value of the current pixel after background removal and noise reduction. The gray value of each pixel after background removal and noise reduction is stored in the register, and the register group transmits the gray value after background removal and noise reduction to Sub-area judgment and extraction module. The line buffer group and register group in the background fluctuation intensity acquisition module extract the current processing sub-region from the fluorescence map, and obtain the local standard deviation of the current processing sub-region, and use the local standard deviation of the current processing sub-region as the background fluctuation of the current processing sub-region Intensity, the background fluctuation intensity of the currently processed sub-area is stored in the register, and transmitted to the sub-area judgment and extraction module. A row of buffer groups and register groups in the sub-region judgment and extraction module extract the current processing sub-region from the image after background removal and noise reduction, and output the sub-region extraction control signal through the sub-region judgment circuit, and control the sub-region extraction The signal is transmitted to the control terminal of the sub-region extraction circuit, another line buffer group and another register group in the sub-region judgment and extraction module extract the candidate sub-region from the fluorescence map, and transmit the candidate sub-region to the sub-region extraction circuit, the sub-region extraction circuit determines in real time whether the candidate sub-region is a fluorescent molecule sub-region according to the sub-region extraction control signal, if so, stores the candidate sub-region into the sub-region data FIFO, and merges the sub-region data FIFO, Cut the merged sub-region data to obtain sub-region data of any size, summarize the region data of the fluorescent molecules output by the two image preprocessors, and transmit them to the positioning processing unit and image reconstruction unit through USB3.0 to process. Because the interface where the sCOMS camera is connected to the FPGA is a camera link interface, the data transmitted by this interface to the FPGA per clock is 5 pixels, which is 5*16bit data volume. Therefore, each image preprocessor consists of five noise reduction processing modules, five background removal processing modules, five background fluctuation intensity acquisition modules, and five sub-region judgment and extraction modules.

Due to the sparse excitation characteristics of super-resolution positioning imaging, only the area containing fluorescent molecules is effective. Only extracting and transmitting the data of the area where these fluorescent molecules are located will greatly reduce the amount of data transmitted and stored, and greatly simplify the workload of subsequent processing. . Fluorescent molecule sub-region data is transmitted to PC through USB3.0 interface. Since the method of obtaining the background fluctuation intensity of each pixel by using the local standard deviation and using it as a threshold is adaptive, it has the advantages of being simple and easy to use without requiring the user to input parameters.

The data processing device for super-resolution positioning microscopic imaging provided by the present invention can reduce the image data of 800MB/s to below 300MB/s. The data of the area where the fluorescent molecules are located is conveniently and effectively transmitted to the user's computer through the USB3.0 interface, and the graphics acceleration processing is realized through the GPU of the user's computer, so as to achieve real-time reconstruction of super-resolution images and realize what you see is what you get.

The data processing method for super-resolution positioning microscopic imaging provided by the present invention comprises the following steps:

(1) Perform Gaussian low-pass filter processing on the fluorescence image collected by the optical signal detector to reduce the noise signal in the fluorescence image, and output the first image. When selecting a 5×5 area centered on the current processing pixel for Gaussian low-pass filtering, the convolution template is:

The first image is subjected to ring filtering processing, the background noise in the first image is removed, and the second image is output. When selecting a 7×7 area centered on the current processing pixel for ring filtering, the convolution template is:

The third current processing sub-region is extracted from the fluorescence image collected by the optical signal detector, and the local standard deviation filtering process is performed on the third current processing sub-region to obtain the local standard deviation of the third current processing sub-region, and the third current processing sub-region is obtained. The local standard deviation of the processing sub-region is used as the background fluctuation intensity of the third currently processing sub-region.

If the 7×7 area with the current processing pixel as the center is selected as the third current processing sub-area, as shown in Figure 6, the 24 outermost pixels are divided into 4 areas, and then the two areas of area 1 and area 2 The gray values of the 12 pixels in the area are sorted from small to large, and the first 6 pixels are selected, and the gray values of the 12 pixels in the two areas of area 3 and area 4 are sorted from small to large, and compared one by one. 6 pixels, and finally 12 pixels are screened out.

According to the formula

Compute the local standard deviation for the third currently processed subregion.

In the formula, x _i is the gray value of the i-th pixel finally screened out,

(2) Extract all fluorescent molecule sub-regions in the fluorescence image according to the fluorescence image, the second image and the background fluctuation intensity, including:

(21) extracting the fourth current processing sub-region from the second image, when the central pixel of the third current processing sub-region in the fluorescent image has the same position as the central pixel of the fourth current processing sub-region in the second image, then Using the background fluctuation intensity of the third current processing sub-region as the background fluctuation intensity of the fourth current processing sub-region;

If the gray value of the center pixel of the fourth currently processed sub-area is the maximum gray value of all pixels in the fourth currently processed sub-area;

And the gray value of the central pixel of the fourth current processing sub-region is greater than twice the background fluctuation intensity of the fourth current processing sub-region;

And the sum of the gray value of the central pixel of the fourth current processing sub-region and the gray value of four neighboring pixels of the central pixel of the fourth current processing sub-region is greater than 9 times of the background fluctuation intensity of the fourth current processing sub-region;

And the sum of the gray value of the central pixel of the fourth current processing sub-region and the gray value of the eight neighboring pixels of the central pixel of the fourth current processing sub-region is greater than 11 times of the background fluctuation intensity of the fourth current processing sub-region;

Then extract the region with the same position in the second image as the fluorescent molecule sub-region in the fluorescence image and the fourth current processing sub-region, and enter step (22); otherwise do not extract the fourth current processing sub-region in the fluorescence image and the fourth current processing sub-region The regions with the same position in the two images,

The 4 neighborhoods of the center pixel include 1 pixel to the left of the center pixel, 1 pixel to the right of the center pixel, 1 pixel above the center pixel, and 1 pixel below the center pixel; the 8 neighborhoods of the center pixel include 1 pixel to the left of the center pixel 2 pixels, 2 pixels to the right of the center pixel, 2 pixels above the center pixel, and 2 pixels below the center pixel;

The fourth currently processed sub-area is a sub-area larger than 7×7 extracted from the second image.

(22) judge whether all the 4th current processing subregions have been extracted, if so, then terminate; otherwise, enter step (21);

(3) Obtain a super-resolution reconstructed image according to the localization processing of the fluorescent molecule sub-regions.

The data preprocessing method for super-resolution positioning microscopic imaging provided by the present invention, the data preprocessing method for super-resolution positioning microscopic imaging provided by the present invention, will extract each third current processing element from the fluorescence map Regional local standard deviation processing, obtaining the background fluctuation intensity of the third current processing sub-region, and extracting the fourth current processing sub-region from the second image, using the background fluctuation intensity of the third current processing sub-region as the fourth current processing sub-region , and determine whether there are fluorescent molecules in the fourth current processing sub-region according to the background fluctuation intensity of the fourth current processing sub-region and the fourth current processing sub-region, which can make the entire image have an uneven background. The extracted fluorescent molecular sub-regions are more accurate, have better adaptability, and obtain accurate super-resolution reconstruction images. Extract the fourth current processing sub-region from the second image, compare the gray value of each pixel in the fourth current processing sub-region with the background fluctuation intensity of the fourth current processing sub-region, and determine whether the fourth current processing sub-region is fluorescent molecule, by judging whether the gray value of the center pixel of the fourth current processing sub-region is the maximum value to determine whether the fluorescent molecule is in the center of the fourth current processing sub-region, the area where the fluorescent molecule is located in the fluorescence map can be accurately judged. Accurate super-resolution reconstruction images are obtained according to the positioning processing of the fluorescent molecules.

Fig. 7 is a schematic diagram of the effect of the fluorescence image preprocessing provided by the data preprocessing circuit provided by the present invention, Fig. 7 (a) is a schematic diagram of the effect of the fluorescence image collected by the sCOMS camera, and Fig. 7 (b) is the noise reduction processing and the fluorescence image Schematic diagram of the effect of background removal processing, Figure 7(c) is a schematic diagram of the effect of background fluctuation intensity after the local standard deviation filtering of the fluorescence image, Figure 7(d) is a schematic diagram of the effect of all fluorescent molecular sub-regions extracted from the fluorescence image .

Those skilled in the art can easily understand that the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, All should be included within the protection scope of the present invention.

Claims (7)

1. a kind of data processing equipment for super-resolution positioning micro-imaging characterized by comprising data prediction circuit And first processor；

The data prediction circuit includes:

Prober interface circuit, for acquiring and transmitting fluorogram；

FPGA, input terminal is connect with the output end of the prober interface circuit, for extracting fluorescent molecule from fluorogram Subregion；And first interface circuit, input terminal are connect with the output end of the FPGA, for fluorescent molecule subregion to be passed Transport to first processor；

The first processor includes:

Memory, first end are connect with the output end of first interface circuit, for storing fluorescent molecule subregion；

CPU for sending positioning instruction to GPU, and receives the Super-resolution Reconstruction figure of GPU output；

GPU, first end are connect with the first end of the CPU, the second end connection of second end and memory, for according to institute It states positioning instruction and localization process is carried out to fluorescent molecule subregion, obtain Super-resolution Reconstruction figure, and Super-resolution Reconstruction figure is transmitted To the CPU, the Super-resolution Reconstruction figure is shown by final image display；

The FPGA includes:

Fluorogram read module (5), input terminal are connect with the output end of prober interface circuit, for obtaining fluorogram, and Fluorogram is divided into the output of three road fluorograms；

Noise reduction process module (1), input terminal are connect with the first output end of the fluorogram read module (5), for glimmering Light figure carries out noise reduction process, exports the first image；

Go to background processing module (2), input terminal is connect with the output end of noise reduction process module (1), for the first image into Row goes background process, exports the second image；

Background fluctuations intensity obtains module (3), including third subregion obtains circuit (301), and input terminal and fluorogram are read Module (5) second output terminal connection, for extracting the currently processed subregion of third from fluorogram；Background fluctuations intensity obtains Circuit (302), input terminal is connect with the output end that third subregion obtains circuit (301), currently processed for obtaining third The Local standard deviation of subregion, and using the Local standard deviation of the currently processed subregion of third as the currently processed subregion of third Background fluctuations intensity, the threshold value as the judgement of fluorescent molecule subregion；And

Subregion judgement and extraction module (4), including the 4th subregion obtain circuit (401), input terminal and go background process The output end of module (2) connects, for extracting the 4th currently processed subregion from the second image；Subregion decision circuitry (402), first input end is connect with the output end that the 4th subregion obtains circuit (401), the second input terminal and background wave Fatigue resistance obtains the output end connection of circuit (302), and the background fluctuations intensity of the currently processed subregion of third is worked as the 4th The background fluctuations intensity of pre-treatment sub-district, according to the background fluctuations intensity and the 4th currently processed son of the 4th currently processed subregion The signal strength in region determines that the 4th currently processed subregion whether there is fluorescent molecule, and exports subregion and extract control letter Number；Alternative subregion obtains circuit (403), and input terminal connect with fluorogram read module (5) the third output end, is used for Alternative subregion is extracted from fluorogram；Subregion extracts circuit (404), and input terminal and alternative subregion extract circuit (403) output end connection, control terminal connect with subregion decision circuitry (402) output end, are extracted and controlled according to subregion Signal determines whether alternative subregion is fluorescent molecule subregion.

2. data processing equipment as described in claim 1, which is characterized in that in data prediction circuit further include:

Multiplex electronics, input terminal are connect with prober interface circuit output end, and the first output end is with the FPGA's Input terminal connection, second output terminal are used to for fluorogram being transmitted to second processor, and multiplex electronics are used to receive and will Fluorogram copies as the output of two-way fluorogram.

3. data processing equipment as described in claim 1, which is characterized in that in data prediction circuit further include:

Multiplex electronics, input terminal are connect with prober interface circuit output end, and the first output end is with the FPGA's Input terminal connection, second output terminal are used to fluorogram being transmitted to second processor, and third end to N-terminal is used to conduct Expansion interface, multiplex electronics are for receiving and fluorogram being copied as the output of multichannel fluorogram, wherein N >=3.

4. data processing equipment as described in claim 1, which is characterized in that the background fluctuations intensity obtains circuit (302) Include:

Pixel comparator (3021), input terminal are connect with the output end that the third subregion obtains circuit (301), are used for root Screening pixel is obtained from the currently processed subregion of third according to the gray value of each pixel of the currently processed subregion of third；

Mean value adder (3022), input terminal are connect with the output end of the pixel comparator (3021), for being screened The average value of pixel；

Subtracter (3023), first input end are connect with the output end of the pixel comparator (3021), the second input terminal It is connect with the output end of mean value adder (3022), for obtaining the difference of each screening pixel and the average value for screening pixel；With And

Local standard deviation adder (3024), input terminal are connect with the output end of subtracter (3023), are used for each screening picture The difference of element and the average value of screening pixel is summed, and the Local standard deviation of the currently processed subregion of third is obtained.

5. data processing equipment as described in claim 1, which is characterized in that the subregion decision circuitry (402) includes:

First pixel comparator (4021), input terminal are connect with the output end that the 4th subregion obtains circuit (401), are used In the center pixel gray value for judging the 4th currently processed subregion whether be the 4th currently processed each grey scale pixel value of subregion Middle maximum value, and the first level value is exported according to judging result；

Second pixel comparator (4022), one input end obtain circuit (401) output end with the 4th subregion and connect, Another input terminal obtains circuit (302) output end with the background fluctuations intensity and connect, for according to the 4th currently processed sub-district The background fluctuations intensity of the currently processed subregion of center pixel sum of the grayscale values the 4th in domain exports second electrical level value；

Four neighborhood pixel addition devices (4023), input terminal obtain circuit (401) output end with the 4th subregion and connect, use In by the middle imago of four neighborhood grey scale pixel values and the 4th currently processed subregion of the 4th currently processed subregion center pixel Plain gray value carries out accumulation process and exports the first cumulative gray value；

Eight neighborhood pixel addition device (4025), input terminal obtain circuit (401) output end with the 4th subregion and connect, use In by the middle imago of the eight neighborhood grey scale pixel value of the 4th currently processed subregion center pixel and the 4th currently processed subregion Plain gray value carries out accumulation process and obtains the second cumulative gray value；

Four neighborhood pixel ratios are compared with device (4024), the output end company of one input end and the four neighborhoods pixel addition device (4023) It connects, another input terminal obtains circuit (302) output end with the background fluctuations intensity and connect, for according to the first cumulative gray scale Value and the background fluctuations intensity of the 4th currently processed subregion export third level value；

Eight neighborhood pixel comparator (4026), one input end are connect with eight neighborhood pixel addition device (4025) output end, Its another input terminal obtains circuit (302) output end with the background fluctuations intensity and connect, for according to the second cumulative gray value The 4th level value is exported with the background fluctuations intensity of the 4th currently processed subregion；And

Logical AND gate (4027), first input end are connect with the output end of the first pixel comparator (4021), and second Input terminal is connect with the output end of the second pixel comparator (4022), third input terminal and the four neighborhoods pixel ratio compared with The output end of device (4024) connects, and the 4th input terminal is connect with the output end of the eight neighborhood pixel comparator (4026), uses Control signal is extracted in exporting subregion according to the first level value to the 4th level value.

6. a kind of data processing method based on any data processing equipment of claim 1 to 5, which is characterized in that packet Include following steps:

(1) noise reduction sonication is carried out to fluorogram and exports the first image, and ambient noise is carried out to the first image and is handled, exported Second image；

The currently processed subregion of third is extracted from fluorogram, and the currently processed subregion of third is carried out at Local standard deviation Reason, obtain the currently processed subregion of third Local standard deviation, and using the Local standard deviation of the currently processed subregion of third as The background fluctuations intensity of the currently processed subregion of third；

(2) the 4th currently processed subregion is extracted from the second image, by the background fluctuations intensity of the currently processed subregion of third It is current according to the signal strength of the 4th currently processed subregion and the 4th as the background fluctuations intensity of the 4th currently processed sub-district The background fluctuations intensity of processing subregion judge that the 4th currently processed subregion whether there is fluorescent molecule subregion, and according to sentencing Disconnected result determine whether using in fluorogram with the 4th currently processed subregion in the second image the identical region in position as glimmering Optical molecule subregion；The 4th currently processed subregion is updated, all fluorescent molecule subregions in fluorogram are extracted；

(3) fluorescent molecule subregion is subjected to localization process, obtains Super-resolution Reconstruction figure.

7. data processing method as claimed in claim 6, which is characterized in that include the following steps: in the step (2)

(21) the 4th currently processed subregion is extracted from the second image, if the ash of the center pixel of the 4th currently processed subregion Angle value is the maximum value in the gray value of the 4th currently processed subregion all pixels,

And the 4th currently processed subregion center pixel gray value be greater than the 4th currently processed subregion background fluctuations it is strong 2 times of degree；

And the 4th currently processed subregion center pixel gray value and the 4th currently processed subregion center pixel four neighborhoods The sum of gray value of pixel is greater than 9 times of the background fluctuations intensity of the 4th currently processed subregion；

And the 4th currently processed subregion center pixel gray value and the 4th currently processed subregion center pixel it is eight adjacent The sum of gray value of domain pixel is greater than 11 times of the background fluctuations intensity of the 4th currently processed subregion；

Then using in fluorogram with the 4th currently processed subregion in the second image the identical region in position as fluorescent molecule Extracted region, and enter step (22)；Otherwise the position in the second image is not extracted in fluorogram with the 4th currently processed subregion Identical region is set, and enters step (22)

(22) judge whether all 4th currently processed subregions have all been extracted, if so, terminating；Otherwise, it enters step (21)；

4th currently processed subregion is the subregion greater than 7 × 7 extracted from the second image.