Summary of the invention
Purpose of the present invention overcomes above-mentioned the deficiencies in the prior art exactly, and a kind of Hadamard transform modulation method of detection sensitivity of array photoelectric detector is provided.This method is utilized the multichannel characteristic of array detector, and the method by Adama coding, decoding can promote the sensitivity of testing result significantly.
Technical scheme provided by the invention is: the Hadamard transform modulation method that improves detection sensitivity of array photoelectric detector, use array photodetectors as detecting device, by the detector acquisition light signal, the cataloged procedure of simulation Hadamard transform addressable part, handle the light signal that photoelectric detector obtains by Hadamard transform encryption algorithm and decoding algorithm, thereby improve detection sensitivity.
Above-mentioned array photodetectors is the photoelectric detector of a linear array or face battle array.
The encryption algorithm of described Hadamard transform and decoding algorithm: carry out opto-electronic conversion as quick unit as the perform region so that the n on the array photodetectors is individual, obtain original light signal ordered series of numbers X
i, i=1 wherein, 2 ... n; With X
iWith Hadamard transform coding templet sequence S
I, jCorrespondence multiplies each other, and again this array summation is obtained y
j, promptly
Y in the formula
jBe the summation of the signal that produces with the modulation of j template coded sequence, j=1 wherein, 2 ... n; Utilize Y=(y
1, y
2Y
n) solving X by the Adama decoding algorithm, X is by the light signal after Hadamard transform encryption algorithm and the decoding algorithm processing.
Array photodetectors of the present invention can be divided into one or more perform regions as quick unit, unallocatedly advance being rejected of perform region as quick unit; For being divided into a plurality of perform regions, all handle according to the encryption algorithm and the decoding algorithm of described Hadamard transform each perform region.
The present invention also provides the equipment of realizing above-mentioned modulator approach, comprises as the array of detectors photodetector, is used to realize that Hadamard transform encryption algorithm and decoding algorithm handle the device of the light signal that photoelectric detector obtains.
The present invention uses array photodetectors as detecting device, by the signal ordered series of numbers X that detecting device is read
iCoded sequence S with the Hadamard transform coding templet
iCorrespondence multiplies each other, and promptly is equal to the projection signal's intensity that obtains mechanical template modulation, this array is sued for peace again, and is to be equal to the y that single-point detector (as photomultiplier) is read
j
Utilize the Adama encryption algorithm to handle the signal that uses detector array to measure, in identical photoelectric conversion time, can provide higher signal to noise ratio (S/N ratio).Promptly count the size (n is the natural number greater than 1, and the present invention can set size and the corresponding n rank S encoder matrix of n as required) of n according to the middle code symbols of the Adama encryption algorithm that uses, signal to noise ratio (S/N ratio) improves (n+1)/2n
1/2Doubly.And do not need the own coding parts, applicability is very extensive.
Utilize this method that the signal modulating-coding of sensor array is handled, may be bigger because of the scale of array, save coding, decode time and need means by " subregion " that some code element is unified to handle.
The coded system that the present invention does not need conventional Hadamard transform to use, but with array detecting device self as coding templet, therefore can be directly used in the existing array detector system.Therefore its advantage is: 1) saved some parts, simplified the circuit of light path and control section; 2) improved coding rate; 3) avoided by adding the measuring error that many seams diffraction that addressable part brings causes.
Embodiment
A kind of coding method that promotes at the signal to noise ratio (S/N ratio) of array-based photodetector that the present invention relates to.Concrete embodiment is that sensor array is gathered a secondary data, and these data and coded sequence multiply each other and encode, and obtains the signal value behind the coding.According to the coding progression of setting, use each coded sequence to handle the data that primary transducer collects successively.Sensor acquisition to data do not reuse.After all coded sequences are finished using, can obtain the data matrix that is used to decode.Utilize the signal that the Adama decode procedure can solve needs mensuration.
Embodiment 1: the Adama coding to the direct-reading spectrometer of the line array CCD that uses 1024 pixels is transformed.
Line array CCD is 1024 pixels, for mating the two-value circulation s-matrix coding that is suitable for n=1023, gives up the pixel in end, uses 1023 pixels wherein to encode.The line array CCD exposure is once multiplied each other by data with the delegation in the 1023 rank s-matrix, is that 1 corresponding measured value keeps behind coding in the S coded sequence, is that 0 corresponding measured value just is 0, and the signal value that this encoding process is crossed adds up, and is a coded data.Same operation is carried out 1023 times, and the data that the n time exposure gathered are used the capable coding of s-matrix n, can obtain the array of 1023 elements.Utilize this array to ask the spectrum that detects by the Adama decoding algorithm.
With 1023 on the array photodetectors as quick unit as the perform region image data and encode, the signal Y that at every turn measures can be write as (1) formula:
(1) y in the formula
jBe summation with the signal of j template coded sequence modulation generation, x
iBe that the last i of this CCD is individual as the signal that quick unit produced, vector S
j=(S
1, j, S
2, j... S
N, j) value be 1 to each element that " sees through ", be 0 to the element of " do not see through ", in order to reduce 1023 all signals, then must carry out 1023 times and measure, equation (2):
Utilize Y=(y
1, y
2Y
n) promptly solve X by the Adama decoding algorithm, just ask and calculate required this n the detected signal in the quick unit of picture.
The resulting data of line array CCD are the spectrum that direct-reading spectrometer reads.Utilize Hadamard transform can the raising system to survey the signal to noise ratio (S/N ratio) of faint spectrum.By 1023 rank Hadamard transforms, the signal to noise ratio (S/N ratio) of spectrographic detection improves (1023+1)/(2 * 1023
1/2)=16.0 times.Detection for low light level spectral signal is more favourable.
The encryption algorithm of above-mentioned Hadamard transform and decoding algorithm can be realized by soft, the hardware of computing machine.
Above-mentioned modulator approach can be by comprising as the array of detectors photodetector, being used to realize that Hadamard transform encryption algorithm and decoding algorithm handle the equipment of the device of the light signal that photoelectric detector obtains and realize.
Embodiment 2: to the encoding process of the area array CCD detector image-forming picture that uses 512 * 512 pixel resolutions.
The resulting data of area array CCD are actually one 512 * 512 data matrix.Utilize Hadamard transform can improve the signal to noise ratio (S/N ratio) of image, but just should not use this time 512 * 512=262144 rank s-matrix to encode, such minute length will be unacceptable.Therefore should use the lower-order s-matrix to encode to image.
This area array CCD can be considered as the set of 512 512 linear arrays.By each linear array is carried out encoding process, also can realize improving the effect of image resolution ratio.With reference to the method for embodiment 1,512 rank square matrixes are given up a column data of afterbody.But this column data can carry out one time 511 coding in addition, does not participate in coded data and has only one.511 data of each row all with 511 rank s-matrix in delegation pursue data and multiply each other, be that 1 corresponding measured value keeps behind coding in the coded sequence, be that 0 corresponding measured value just is 0, the signal value that this encoding process is crossed adds up, and is the coding array of one 511 element.Same operation is carried out 511 times, and the data that the n time exposure gathered are used the capable coding of s-matrix n, can obtain the ordered series of numbers group of 512 511 elements.Utilize this ordered series of numbers group of 511 * 512 to decode, just can obtain 512 * 511 data matrix, one 511 ordered series of numbers of adding extra delegation is by coding, decoded results, and a pixel that has neither part nor lot in coding, promptly can obtain 512 * 512 data matrix.Utilize this data matrix can obtain the square-shaped image of the length of side 512 pixels, the signal to noise ratio (S/N ratio) of picture improves (511+1)/(2 * 511
1/2)=11.3 times (influence that has neither part nor lot in data of coding can be ignored).
Embodiment 3: to the encoding process of the area array CCD detector image-forming picture that uses 600 * 400 pixel resolutions.
The resulting data of area array CCD are actually one 600 * 400 data matrix.Utilize Hadamard transform can improve the signal to noise ratio (S/N ratio) of image, similar with embodiment 2, should not use 600 * 400=240000 rank S coded sequence to encode.Therefore should use lower-order S coded sequence to encode to image.
This area array CCD can be divided into several zonules of 15 * 17.By each zone being utilized 255 rank S circular matrix encoding process, also can realize improving the effect of image resolution ratio.Because the sensor pixel ratio is not 15 * 17 integral multiple, so the data at edge need be cast out.There is the zone of 9 row, 400 pixels to be rejected.
Each zone by the decoding processing on 255 rank, can obtain the image of one 600 * 391 pixel after carrying out encoding process by 255 codings.The signal to noise ratio (S/N ratio) of this picture improves (255+1)/(2 * 255
1/2)=8.0 times.