CN114295108A - A ranging method and system for external equipment and infrared telescopes - Google Patents

Abstract

The invention relates to a distance measuring method and a distance measuring system based on external equipment and an infrared telescope, infrared glass lenses with different curvatures and different refractive indexes are arranged on the same axis in the infrared telescope to form an infrared imaging lens, an infrared photosensitive unit is arranged into an uncooled infrared focal plane array to form an infrared detector in the infrared telescope, and the distance measuring method of the infrared telescope comprises the following steps: determining the category of a target to be detected, calibrating a data set, training a detection model, importing the model, detecting the target, measuring the distance of the target, identifying an image and observing and displaying. The method comprises the steps of acquiring an infrared data set, calibrating the data set and training a model to obtain a detection model for target detection, and performing category identification, target ranging and image identification on an infrared target by using the detection model and then performing identification superposition to complete real-time ranging and labeling of the infrared image target.

Description

A ranging method and system for external equipment and infrared telescopes

technical field

The invention relates to a ranging system, in particular to a ranging method and system based on an external device and an infrared telescope.

Background technique

An infrared camera is an electronic device that receives infrared light emitted or reflected by the scene itself and reaches the purpose of imaging the scene. The infrared telescope is an observation product that integrates a miniaturized infrared camera, an electronic display screen and a battery module, and can transmit an instant infrared image to the user. Because infrared images can distinguish targets of different temperatures and are not affected by ambient light, they are often used in fields such as field hunting and police reconnaissance, especially at night. When the user needs to obtain the distance value of the observed target, two methods of physical ranging and image ranging can be used. Physical ranging is to integrate an optical ranging module on the infrared telescope, and obtain the distance of the target by calculating the time difference between the emitted beam and the reflected beam. Physical ranging will increase the size, weight, power consumption and cost of the infrared telescope. , so image ranging is more advantageous; image ranging is a method of directly obtaining the target distance by using infrared images without the aid of other equipment. Common image ranging methods are divided into manual measurement method and grid comparison method. The manual measurement method is to manually mark the upper and lower boundaries of the target on the screen, and input the type or physical height value of the target, and the system can calculate the distance value of the target; the grid comparison method is to superimpose horizontal or vertical on the electronic display screen. After selecting the type of target, the size of the scale value occupied by the target at different distances is displayed on the screen, and the estimated value of the target distance can be obtained by comparing the actual scale value occupied by the target with the theoretical value.

The grid comparison method cannot accurately obtain the distance value of the target, and the error of distance estimation is large; the manual measurement method is complicated in operation and lacks real-time performance, especially for moving targets, which is very difficult to measure.

SUMMARY OF THE INVENTION

The technical problem solved by the invention is: constructing an infrared telescope ranging method and system to overcome the technical problems of complicated ranging operation and large ranging error of the infrared telescope in the prior art.

The technical scheme of the present invention is as follows: constructing a distance measuring method for an infrared telescope, infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens, and the infrared photosensitive unit is set as an uncooled infrared focal plane The array forms an infrared detector in an infrared telescope, which is characterized by comprising the following steps:

Model import: for the infrared telescope embedded hardware platform used, after the training detection model is converted into a format matching the embedded platform, the infrared telescope embedded system imports the training detection model; the training detection model is based on the target shape characteristics and The infrared image of the target is established and formed in an external device;

Target detection: After the infrared telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target;

Target ranging: Calculate the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the known focal length of the infrared telescope and the pixel size of the detector;

Image identification: obtain the current target distance to form the target distance identification;

Observation display: The image on the electronic display screen is projected onto the retina of the human eye through an eyepiece formed by arranging infrared and visible glass lenses with different shapes and different refractive indices.

A further technical solution of the present invention is: according to the preset height h ₀ of the target to be measured, the pixel height of the target person or object in the picture is q, the focal length f of the lens of the infrared telescope and the size d of the detector pixel are known, which can be obtained by The following formula calculates the distance L between the target person or object and the observer:

A further technical solution of the present invention is: forming the category identification of the acquired current target category, and superimposing the target category identification and the distance identification on the infrared image in the form of an on-screen menu.

The technical scheme of the present invention is to provide an infrared telescope ranging method based on external equipment training,

Complete the establishment of the training detection model in the external device, including the following steps:

Determine the category of the target to be measured: classify the target to be measured according to the shape characteristics of the target;

Infrared image data acquisition: The infrared image of the target to be measured is obtained by shooting the infrared image shooting equipment, and the infrared image data collection is formed by various types of target images;

Data set calibration: After obtaining the infrared data set, the calibration of the target to be measured in the infrared image is completed. The calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target;

Training detection model: establish a detection model, use the calibrated data as input data, and obtain complete model parameters;

Ranging in an infrared telescope includes the following steps:

Model import: For the infrared telescope embedded hardware platform used, after the training detection model is converted into a format matching the embedded platform, the infrared telescope embedded system imports the training detection model;

Target detection: After the infrared telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target;

Target ranging: Calculate the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the known focal length of the infrared telescope and the pixel size of the detector;

Image identification: obtain the current target distance to form the target distance identification;

Observation display: The image on the electronic display screen is projected onto the retina of the human eye through an eyepiece formed by arranging infrared and visible glass lenses with different shapes and different refractive indices.

A further technical solution of the present invention is that: the infrared image in the infrared image data acquisition is divided into a sample infrared image and a background infrared image.

The further technical scheme of the present invention is: after the data set is calibrated, the infrared data set is augmented, and a new infrared data sample is generated by the original infrared data sample, and the infrared data sample augmentation includes single image processing and multi-image processing.

The technical scheme of the present invention is to construct an infrared telescope ranging system, including an infrared imaging lens, an infrared detector, an embedded processor module, an electronic display screen, an eyepiece module, and infrared glass lenses with different curvatures and different refractive indices in the infrared The infrared imaging lens is arranged on the same axis in the telescope, and the infrared photosensitive unit is arranged as an uncooled infrared focal plane array to form an infrared detector in the infrared telescope. The embedded processor module includes a target detection module, a target ranging module, an image The identification module, for the infrared telescope embedded hardware platform used, converts the training detection model into a format matching the embedded platform, and imports the training detection model into the infrared telescope embedded system; the training detection model is based on the target shape characteristics and The infrared image of the target is established and formed by external equipment. After the infrared telescope obtains the infrared image, it enters the training detection model. The target detection module obtains the target category in the infrared image and the rectangular frame covering the entire target. The target ranging module According to the preset height of the target to be measured, the pixel height of the target person or object in the picture, the known focal length of the infrared telescope and the pixel size of the detector, the distance between the target and the observer is calculated, and the image identification module The current target distance is obtained to form a target distance identifier, and an eyepiece module is formed by arranging infrared and visible glass lenses with different shapes and different refractive indices. The eyepiece module projects the image on the electronic display screen onto the retina of the human eye.

A further technical solution of the present invention is that the eyepiece module changes the position of one or more visible light glass lenses through the lens position adjustment mechanism, thereby changing the virtual image position of the electronic display screen.

A further technical solution of the present invention is: further comprising an interface management unit, the interface management unit manages the reception of infrared detector signals, the reception of I/O signals of the key module, and the data output to the electronic display screen.

A further technical solution of the present invention is: further comprising a non-uniform correction module, the non-uniform correction module adopts an embedded processor module to perform non-uniform correction on the array digital signal of the infrared detector to correct the photosensitive unit of the infrared detector temperature response rate.

The technical scheme of the present invention is to construct an infrared telescope ranging system based on external equipment, which is characterized in that it includes external equipment and an infrared telescope, and the external equipment includes a training detection model generation module, and the training detection model generation module is in In the process of generating the detection model, the target to be tested is classified according to the shape characteristics of the target, and the infrared image of the target to be tested is obtained by shooting with an infrared image shooting device, and various types of target images are formed into infrared image data collection, and after obtaining the infrared data set, The calibration of the target to be measured in the infrared image is completed. The calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target. A detection model is established, and the calibrated data is used as input data to obtain complete model parameters; The infrared telescope includes: an infrared imaging lens, an infrared detector, an embedded processor module, an electronic display screen, an eyepiece module, and infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens, The infrared photosensitive unit is arranged as an uncooled infrared focal plane array to form an infrared detector in the infrared telescope, and the embedded processor module includes a target detection module, a target ranging module, and an image identification module. Platform, after converting the training detection model into a format matching the embedded platform, the embedded system of the infrared telescope imports the training detection model, after the infrared telescope obtains the infrared image, it enters the training detection model, and the target detection module obtains the infrared image The target category and the rectangular frame covering the entire target, the target ranging module is based on the preset height of the target to be measured, the pixel height of the target person or object in the picture, the known focal length of the infrared telescope and the detector pixel. Size, calculate the distance between the target and the observer, the image identification module obtains the current target distance to form the target distance identification, and the eyepiece module formed by arranging infrared and visible glass lenses with different shapes and different refractive indices, the eyepiece module will electronically display The image on the screen is projected onto the retina of the human eye.

The technical effect of the invention is as follows: constructing a distance measuring method and system for an infrared telescope based on external equipment training, infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens, an infrared photosensitive unit Set up an uncooled infrared focal plane array to form an infrared detector in an infrared telescope, and complete the establishment of a training detection model in an external device, including the following steps: determining the category of the target to be measured: classifying the target to be measured according to the shape characteristics of the target; infrared Image data acquisition: The infrared image of the target to be measured is obtained by taking infrared image shooting equipment, and various types of target images are formed into infrared image data collection; Calibration, the calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target; training the detection model: establish a detection model, use the calibrated data as input data, and obtain complete model parameters; test in an infrared telescope. It includes the following steps: model import: for the infrared telescope embedded hardware platform used, after converting the training detection model into a format matching the embedded platform, the infrared telescope embedded system imports the training detection model; target detection: infrared After the telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target; target ranging: According to the preset height of the target to be measured, the pixels of the target person or object in the screen The height, the focal length of the lens of the infrared telescope and the pixel size of the detector are known, and the distance between the target and the observer is calculated; image identification: obtain the current target distance to form the target distance identification; The eyepiece formed by the arrangement of visible light glass lenses projects the image on the electronic display onto the retina of the human eye. The invention obtains a detection model for target detection by collecting infrared data sets, data set calibration and model training, and uses the detection model to carry out category recognition, target ranging, image identification and identification overlay for infrared targets, so as to complete the real-time detection of infrared image targets. The ranging and labeling solves the technical problems of complex ranging operations and large ranging errors in the prior art.

Description of drawings

FIG. 1 is a schematic structural diagram of an infrared telescope of the present invention.

FIG. 2 is a schematic structural diagram of an infrared telescope and a training detection model generation module of the present invention.

FIG. 3 is a schematic diagram of the structure included in the embedded processor module of the present invention.

Detailed ways

The technical solutions of the present invention will be further described below with reference to specific embodiments.

The technical embodiment of the present invention is as follows: constructing a distance measuring method for an infrared telescope. Infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens, and the infrared photosensitive unit is set to an uncooled infrared focus. The planar array forms an infrared detector in an infrared telescope, including the following steps:

Model import: for the infrared telescope embedded hardware platform used, after the training detection model is converted into a format matching the embedded platform, the infrared telescope embedded system imports the training detection model; the training detection model is based on the target shape characteristics and The infrared image of the target is created by the external device.

The specific implementation is as follows: an image processing chip is used as an embedded hardware platform, the detection model format matched with the embedded platform is wk format, and a standard deep learning framework is required to generate a wk format file, and the framework supported by the image processing chip is Caffe framework (Convolutional Architecture for Fast Feature Embedding convolutional neural network framework), in a specific embodiment, it is necessary to convert the Yolov3 model to the Caffe framework through a model conversion tool such as RuyiStudio application program, and generate the corresponding wk format file under the Caffe framework ; Import the wk file into the memory of the image processing chip and be called by the software in the process of realizing the target detection.

Target detection: After the infrared telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target.

The specific embodiment is as follows: the infrared telescope is aimed at the scene including the target to be identified, after the software enables the target detection function, the wk model file is called from the memory of the embedded chip to generate the detection model, and the software loads the obtained infrared image into the detection model Calculated in the infrared image to obtain the target category in the infrared image and the rectangular frame covering the entire target.

Target ranging: Calculate the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the lens focal length of the known infrared telescope and the size of the detector pixel.

The specific embodiment is: the pixel height of the rectangular frame covering the entire target obtained by target detection is regarded as the pixel height of the target person or object in the picture as q, the preset height h ₀ of the target to be measured, the known focal length of the infrared telescope lens f and the detector pixel size d, the distance L between the target person or object and the observer can be calculated by the following formula:

Taking the detection of a target person as an example, the pixel height of the rectangular frame covering the target person obtained by target detection is 100 pixels, the preset height of the person is 1.7m, the focal length of the infrared telescope lens is 20mm, and the pixel size of the detector is 17um. According to the formula, the target person and the Distance between observers:

Image identification: The obtained category of the current target is formed into a category identification, the current target distance is obtained to form a target distance identification, and the target category identification and distance identification are superimposed on the infrared image in the form of an on-screen menu.

The specific implementation is: the category of the target can be displayed in English or Chinese but not limited to, the target distance can be displayed in but not limited to "meter" as the distance unit, and the target category and target distance can be displayed near the target, such as directly above the target The position can also be displayed around the screen, such as the left side of the screen, and the target and the logo are associated by line segments; the logo display area can use a transparent background, a grayscale background or a background of other colors.

Observation display: The image on the electronic display screen is projected onto the retina of the human eye through an eyepiece formed by arranging infrared and visible glass lenses with different shapes and different refractive indices.

The technical embodiment of the present invention is to provide an infrared telescope ranging method based on external equipment training.

The establishment of the training detection model is completed in an external device, and the external device mentioned here may be an external PC computer, a notebook computer, or other processing system device with strong processing capability. It includes the following steps:

Determine the category of the target to be tested: classify the target to be tested according to the target shape features.

The specific implementation process is as follows: classify the objects to be measured according to their shape features, and the categories of objects to be measured refer to object categories with the same physical characteristics, such as people, cats, chairs, etc. The number of object categories to be measured is limited.

Infrared image data collection: The infrared image of the target to be measured is obtained by shooting the infrared image shooting equipment, and the infrared image data collection is formed by various types of target images.

The specific implementation process is as follows: the infrared image in the infrared image data collection is divided into a sample infrared image and a background infrared image. The sample infrared image should contain one or more objects in the target category to be measured, and the background infrared image does not appear. The category of the target to be tested. In the process of infrared image data acquisition, the selection of color modes such as black heat, white heat, and false color should be determined according to the color mode used in the actual detection process, and one or more of these color modes should be selected.

Data set calibration: After obtaining the infrared data set, the calibration of the target to be measured in the infrared image is completed. The calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target.

The specific embodiment is: after obtaining the infrared data set, the calibration of the target to be measured in the infrared image is completed by the operation, and the calibration content includes the category to which each target to be measured belongs and a rectangular frame that can cover the entire target. After the data set is calibrated, the data set is augmented, and new data samples are generated from the original infrared data samples. The sample augmentation includes single image processing and multi-image processing; single image processing is to adjust the brightness, contrast, hue, Saturation and noise, randomly increase scale change, crop, flip and rotate to form a new infrared image, and adjust the corresponding target information text file as a new infrared image sample; multi-image processing is performed by combining two infrared images. The pixel values are weighted, or multiple images are cropped and spliced to the same infrared image to form a new infrared image, and the corresponding target information text file is adjusted as a new infrared image sample.

In a specific embodiment, the operation can be a manual calibration operation, and the rectangular frame formed by it is as small as possible. The "minimum rectangle" that can be manipulated. The specific method is as follows: the position of the upper boundary of the minimum rectangular frame is the uppermost pixel of the object to be measured, the position of the lower boundary is the lowermost pixel of the object to be measured, the position of the left boundary is the position of the leftmost pixel of the object to be measured, and the position of the right boundary is The rightmost pixel of the target to be measured; after calibration, each infrared image corresponds to a text file, and the file saves all the information of the target to be measured in the infrared image. The information of the target to be measured includes the category of the target and the smallest rectangular frame covering the target. The position coordinates in the image, the position coordinates usually take the upper left corner of the infrared image as the coordinate origin, the right direction is the positive direction of the horizontal coordinate, and the downward direction is the positive direction of the vertical coordinate. It can also be determined by the coordinates of the upper left corner and the pixel value of the length and width of the rectangular frame.

Training detection model: Establish a detection model, use the calibrated data as input data, and obtain complete model parameters.

The specific embodiment is: the training detection model is based on the Yolov3 feature extraction network, and after the backbone network Darknet-53, three deep convolutional layers, four maximum pooling layers of different scales, and one composed of 1 and 5 are added. Residual network module composed of deep convolutional layer branches.

Ranging in an infrared telescope includes the following steps:

Model import: For the used embedded hardware platform, convert the model into a format matching the embedded platform and import it into the embedded system.

The specific embodiment is: for the used infrared telescope embedded hardware platform, after the training detection model is converted into a format matching the embedded platform, the embedded system of the infrared telescope imports the training detection model; the training detection model is based on the shape of the target. Infrared images of features and targets are created with external equipment. The image processing chip is used as the embedded hardware platform, and the detection model format that matches the embedded platform is wk format. A standard deep learning framework is required to generate wk format files. The framework supported by the image processing chip is the Caffe framework (Convolutional Architecture for Fast Feature Embedding convolutional neural network framework), in a specific embodiment, it is necessary to convert the Yolov3 model to the Caffe framework through a model conversion tool such as the RuyiStudio application, and generate the corresponding wk format file under the Caffe framework; import the wk file to In the memory of the image processing chip, it is called by the software in the process of realizing the target detection.

Target detection: After the infrared telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target.

The specific embodiment is as follows: the infrared telescope is aimed at the scene including the target to be identified, after the software enables the target detection function, the wk model file is called from the memory of the embedded chip to generate the detection model, and the software loads the obtained infrared image into the detection model Calculated in the infrared image to obtain the target category in the infrared image and the rectangular frame covering the entire target.

Target ranging: Calculate the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the lens focal length of the known infrared telescope and the size of the detector pixel.

The specific embodiment is: the pixel height of the rectangular frame covering the entire target obtained by target detection is regarded as the pixel height of the target person or object in the picture as q, the preset height h ₀ of the target to be measured, the known focal length of the infrared telescope lens f and the detector pixel size d, the distance L between the target person or object and the observer can be calculated by the following formula:

Taking the detection of a target person as an example, the pixel height of the rectangular frame covering the target person obtained by target detection is 100 pixels, the preset height of the person is 1.7m, the focal length of the infrared telescope lens is 20mm, and the pixel size of the detector is 17um. According to the formula, the target person and the Distance between observers:

Image identification: The obtained category of the current target is formed into a category identification, the current target distance is obtained to form a target distance identification, and the target category identification and distance identification are superimposed on the infrared image in the form of an on-screen menu.

The specific implementation is: the category of the target can be displayed in English or Chinese but not limited to, the target distance can be displayed in but not limited to "meter" as the distance unit, and the target category and target distance can be displayed near the target, such as directly above the target The position can also be displayed around the screen, such as the left side of the screen, and the target and the logo are associated by line segments; the logo display area can use a transparent background, a grayscale background or a background of other colors.

Observation display: The image on the electronic display screen is projected onto the retina of the human eye through an eyepiece formed by arranging infrared and visible glass lenses with different shapes and different refractive indices.

As shown in FIG. 2 and FIG. 3 , the specific embodiment of the present invention is to construct an infrared telescope ranging system, including an infrared imaging lens 1, an infrared detector 2, an embedded processor module 3, an electronic display screen 4, an eyepiece Module 5. Infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens 1, and the infrared photosensitive unit is arranged as an uncooled infrared focal plane array to form an infrared detector 2 in the infrared telescope. The embedded processor module 3 includes a target detection module 31, a target ranging module 32, and an image identification module 33. For the infrared telescope embedded hardware platform used, after the training detection model is converted into a format matching the embedded platform, The embedded system of the infrared telescope imports the training detection model. After the infrared telescope obtains the infrared image, it enters the training detection model. The target detection module 31 obtains the target category in the infrared image and the rectangular frame covering the entire target. The distance module 32 calculates the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the lens focal length of the known infrared telescope and the detector pixel size, and the The image identification module 33 obtains the current target distance to form a target distance identification. The eyepiece module 5 is formed by arranging infrared and visible glass lenses with different shapes and different refractive indices. The eyepiece module 5 projects the image on the electronic display screen onto the retina of the human eye.

The specific implementation process of the present invention is as follows: an image processing chip is used as an embedded hardware platform, the detection model format matched with the embedded platform is wk format, and a standard deep learning framework is required to generate a wk format file, and the image processing chip supports The framework is the Caffe framework (Convolutional Architecture for Fast FeatureEmbedding Convolutional Neural Network Framework). In a specific embodiment, it is necessary to convert the Yolov3 model to the Caffe framework through a model conversion tool such as the RuyiStudio application, and generate the corresponding wk under the Caffe framework. Format file; import the wk file into the memory of the image processing chip, and be called by the software in the process of realizing target detection. The infrared telescope is aimed at the scene including the target to be identified. After the software enables the target detection function, the wk model file is called from the memory of the embedded chip to generate the detection model. The software loads the obtained infrared image into the detection model for calculation. The target detection module 31 obtains the target category in the infrared image and the rectangular frame covering the entire target. The pixel height of the rectangular frame covering the entire target obtained by target detection is regarded as the pixel height of the target person or object in the picture as q, the preset height h ₀ of the target to be measured, the known infrared telescope lens focal length f and detector image Element size d, the distance L between the target person or object and the observer can be calculated by the following formula:

Taking the detection of a target person as an example, the pixel height of the rectangular frame covering the target person obtained by target detection is 100 pixels, the preset height of the person is 1.7m, the focal length of the infrared telescope lens is 20mm, and the pixel size of the detector is 17um. According to the formula, the target person and the Distance between observers:

The target ranging module 32 calculates the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the known focal length of the infrared telescope and the pixel size of the detector. distance.

The category of the target can be displayed in English or Chinese but not limited to, the target distance can be but not limited to use "meter" as the distance unit, the target category and target distance can be displayed near the target, such as the position directly above the target, can also be displayed Around the screen, such as the left side of the screen, the target and the logo are associated by line segments; the logo display area can use a transparent background, a grayscale background or a background of other colors. The image identification module 33 obtains the current target distance to form a target distance identification, and the eyepiece module 5 is formed by arranging infrared and visible glass lenses with different shapes and different refractive indices. The eyepiece module 5 projects the image on the electronic display screen onto the retina of the human eye. .

In the process of generating the detection model, the training detection model generation module 6 classifies the target to be measured according to the shape characteristics of the target, obtains the infrared image of the target to be measured by photographing the infrared image capturing device, and forms the infrared image of the target images of various categories. Data acquisition. After obtaining the infrared data set, the calibration of the target to be measured in the infrared image is completed. The calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target. A detection model is established, and the calibrated data is used as input. data to get the complete model parameters;

A preferred embodiment of the present invention is that the eyepiece module 5 changes the position of one or more visible light glass lenses through the lens position adjustment mechanism, thereby changing the virtual image position of the electronic display screen. The eyepiece module may use a mirror.

The preferred embodiment of the present invention further includes an interface management unit, which manages the reception of infrared detector signals, the reception of I/O signals of the key module, and the data output to the electronic display screen. It also includes a non-uniform correction module, which adopts an embedded processor module to perform non-uniform correction on the array digital signal of the infrared detector to correct the temperature responsivity of the infrared detector photosensitive unit.

The technical scheme of the present invention is to construct an infrared telescope ranging system based on external equipment, which is characterized in that it includes external equipment and an infrared telescope, and the external equipment includes a training detection model generation module 6, and the training detection model generation module 6 In the process of generating the detection model, the target to be tested is classified according to the shape characteristics of the target, and the infrared image of the target to be tested is obtained by shooting with an infrared image shooting device, and various types of target images are formed into infrared image data collection, and an infrared data set is obtained. After completing the calibration of the target to be measured in the infrared image, the calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target, establish a detection model, and use the calibrated data as input data to obtain complete model parameters The infrared telescope includes: infrared imaging lens 1, infrared detector 2, embedded processor module 3, electronic display screen 4, eyepiece module 5, the same axis of infrared glass lenses with different curvatures and different refractive indices in the infrared telescope The infrared imaging lens 1 is formed by arranging on the top, and the infrared photosensitive unit is arranged as an uncooled infrared focal plane array to form the infrared detector 2 in the infrared telescope. The embedded processor module 3 includes a target detection module 31, a target ranging module 32, an image The identification module 33, for the used infrared telescope embedded hardware platform, converts the training detection model into a format matching the embedded platform, and imports the training detection model into the infrared telescope embedded system, and after the infrared telescope obtains the infrared image, enters the The training detection model is described, the target detection module 31 obtains the target category in the infrared image and the rectangular frame covering the entire target, and the target ranging module 32 is based on the preset height of the target to be measured, the target person or object in the picture. The pixel height, the focal length of the lens of the known infrared telescope and the size of the detector pixel, calculate the distance between the target and the observer, the image identification module 33 obtains the current target distance to form the target distance mark, and infrared rays with different shapes and different refractive indices The eyepiece module 5 is formed by arranging the visible light glass lenses, and the eyepiece module 5 projects the image on the electronic display screen onto the retina of the human eye.

Complete the establishment of the training detection model in the external device, including the following steps:

The objects to be tested are classified according to the shape features of the objects, and the categories of objects to be tested refer to object categories with the same physical characteristics, such as people, cats, chairs, etc. The number of categories of objects to be tested is limited. The infrared image in the infrared image data collection is divided into a sample infrared image and a background infrared image. The sample infrared image should contain one or more objects in the target category to be measured, and the target category to be detected does not appear in the background infrared image. In the process of infrared image data acquisition, the selection of color modes such as black heat, white heat, and false color should be determined according to the color mode used in the actual detection process, and one or more of these color modes should be selected. After the infrared data set is obtained, the calibration of the target to be measured in the infrared image is completed by the operation, and the calibration content includes the category to which each target to be measured belongs and a rectangular frame that can cover the entire target. After the data set is calibrated, the data set is augmented, and new data samples are generated from the original infrared data samples. The sample augmentation includes single image processing and multi-image processing; single image processing is to adjust the brightness, contrast, hue, Saturation and noise, randomly increase scale change, crop, flip and rotate to form a new infrared image, and adjust the corresponding target information text file as a new infrared image sample; multi-image processing is performed by combining two infrared images. The pixel values are weighted, or multiple images are cropped and spliced to the same infrared image to form a new infrared image, and the corresponding target information text file is adjusted as a new infrared image sample.

In a specific embodiment, the operation can be a manual calibration operation, and the rectangular frame formed by it is as small as possible. The "minimum rectangle" that can be manipulated. The specific method is as follows: the position of the upper boundary of the minimum rectangular frame is the uppermost pixel of the object to be measured, the position of the lower boundary is the lowermost pixel of the object to be measured, the position of the left boundary is the position of the leftmost pixel of the object to be measured, and the position of the right boundary is The rightmost pixel of the target to be measured; after calibration, each infrared image corresponds to a text file, and the file saves all the information of the target to be measured in the infrared image. The information of the target to be measured includes the category of the target and the smallest rectangular frame covering the target. The position coordinates in the image, the position coordinates usually take the upper left corner of the infrared image as the coordinate origin, the right direction is the positive direction of the horizontal coordinate, and the downward direction is the positive direction of the vertical coordinate. It can also be determined by the coordinates of the upper left corner and the pixel value of the length and width of the rectangular frame. A detection model is established, and the calibrated data is used as input data to obtain complete model parameters. The specific embodiment is: the training detection model is based on the Yolov3 feature extraction network, and after the backbone network Darknet-53, three deep convolutional layers, four maximum pooling layers of different scales, and one composed of 1 and 5 are added. Residual network module composed of deep convolutional layer branches.

The technical effect of the invention is as follows: a method and a system for measuring distances of an infrared telescope are constructed. Infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens, and the infrared photosensitive unit is set as an uncooled infrared The focal plane array forms an infrared detector in an infrared telescope. The infrared telescope ranging method includes the following steps: determining the category of the target to be measured: classifying the target to be measured according to the shape characteristics of the target; infrared image data collection: obtained by shooting with infrared image shooting equipment Infrared image of the target to be measured, various types of target images are formed into infrared image data acquisition; data set calibration: after obtaining the infrared data set, the calibration of the target to be measured in the infrared image is completed, and the calibration content includes the belonging of each target to be measured. category and a rectangular box that can cover the entire target; training detection model: establish a detection model, use the calibrated data as input data, and obtain complete model parameters; model import: according to the embedded hardware platform used, convert the model into After matching the format of the embedded platform, import it into the embedded system; target detection: after the infrared telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target; target ranging: According to the preset height of the target to be measured, the pixel height of the target person or object in the picture, the lens focal length of the known infrared telescope and the size of the detector pixel, the distance between the target and the observer is calculated; image identification: obtain the current The target distance forms the target distance identification; Observation display: The image on the electronic display screen is projected onto the retina of the human eye through the eyepiece formed by the arrangement of infrared and visible glass lenses with different shapes and different refractive indices. The invention obtains a detection model for target detection by collecting infrared data sets, data set calibration and model training, and uses the detection model to carry out category recognition, target ranging, image identification and identification overlay for infrared targets, so as to complete the real-time detection of infrared image targets. The ranging and labeling solves the technical problems of complex ranging operations and large ranging errors in the prior art.

The above content is a further detailed description of the present invention in combination with specific preferred embodiments, and it cannot be considered that the specific implementation of the present invention is limited to these descriptions. For those of ordinary skill in the technical field of the present invention, without departing from the concept of the present invention, some simple deductions or substitutions can be made, which should be regarded as belonging to the protection scope of the present invention.

Claims (10)

1. A distance measuring method for an infrared telescope. Infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens, and the infrared photosensitive unit is set as an uncooled infrared focal plane array to form an infrared imaging lens in the infrared telescope. Infrared detector, is characterized in that, comprises the following steps: Model import: for the infrared telescope embedded hardware platform used, after the training detection model is converted into a format matching the embedded platform, the infrared telescope embedded system imports the training detection model; the training detection model is based on the target shape characteristics and The infrared image of the target is established and formed in an external device; Target detection: After the infrared telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target; Target ranging: Calculate the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the known focal length of the infrared telescope and the pixel size of the detector; Image identification: obtain the current target distance to form the target distance identification; Observation display: The image on the electronic display screen is projected onto the retina of the human eye through an eyepiece formed by arranging infrared and visible glass lenses with different shapes and different refractive indices. 2. The infrared telescope ranging method according to claim 1 is characterized in that, according to the preset height h ₀ of the target to be measured, the pixel height of the target person or object in the picture is q, the known lens focal length f of the infrared telescope and the detector pixel size d, the distance L between the target person or object and the observer can be calculated by the following formula:

3 . The infrared telescope ranging method according to claim 1 , wherein the acquired category of the current target is formed into a category identification, and the target category identification and the distance identification are superimposed on the infrared image in the form of an on-screen menu. 4 . 4. an infrared telescope ranging method based on external equipment training, is characterized in that, Complete the establishment of the training detection model in the external device, including the following steps: Determine the category of the target to be measured: classify the target to be measured according to the shape characteristics of the target; Infrared image data acquisition: The infrared image of the target to be measured is obtained by shooting the infrared image shooting equipment, and the infrared image data collection is formed by various types of target images; Data set calibration: After obtaining the infrared data set, the calibration of the target to be measured in the infrared image is completed. The calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target; Training detection model: establish a detection model, use the calibrated data as input data, and obtain complete model parameters; Ranging in an infrared telescope includes the following steps: Model import: For the infrared telescope embedded hardware platform used, after the training detection model is converted into a format matching the embedded platform, the infrared telescope embedded system imports the training detection model; Target detection: After the infrared telescope obtains the infrared image, it enters the training detection model to obtain the target category in the infrared image and the rectangular frame covering the entire target; Target ranging: Calculate the distance between the target and the observer according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the known focal length of the infrared telescope and the pixel size of the detector; Image identification: obtain the current target distance to form the target distance identification; Observation display: The image on the electronic display screen is projected onto the retina of the human eye through an eyepiece formed by arranging infrared and visible glass lenses with different shapes and different refractive indices. 5 . The infrared telescope ranging method based on external equipment training according to claim 4 , wherein the infrared images in the infrared image data collection are divided into sample infrared images and background infrared images. 6 . 6. The infrared telescope ranging method based on external equipment training according to claim 4, is characterized in that, after the data set is calibrated, the infrared data set is augmented, and a new infrared data sample is generated by the original infrared data sample, and the infrared data Sample augmentation includes single-image processing and multi-image processing. 7. An infrared telescope ranging system, characterized in that it comprises an infrared imaging lens, an infrared detector, an embedded processor module, an electronic display screen, an eyepiece module, and infrared glass lenses with different curvatures and different refractive indices in the infrared telescope. The infrared imaging lens is arranged on the same axis of the infrared imaging lens, and the infrared photosensitive unit is arranged as an uncooled infrared focal plane array to form an infrared detector in the infrared telescope. The embedded processor module includes a target detection module, a target ranging module, and an image identification module. , for the infrared telescope embedded hardware platform used, after the training detection model is converted into a format matching the embedded platform, the embedded system of the infrared telescope imports the training detection model; the training detection model is based on the shape characteristics of the target and the The infrared image is formed by the external equipment. After the infrared telescope obtains the infrared image, it enters the training detection model. The target detection module obtains the target category in the infrared image and the rectangular frame covering the entire target. Assuming the height of the target to be measured, the pixel height of the target person or object in the screen, the focal length of the lens of the known infrared telescope and the pixel size of the detector, the distance between the target and the observer is calculated, and the image identification module obtains the current The target distance forms a target distance mark, and an eyepiece module is formed by arranging infrared and visible glass lenses with different shapes and different refractive indices. The eyepiece module projects the image on the electronic display screen onto the retina of the human eye. 8 . The infrared telescope ranging system according to claim 7 , wherein the eyepiece module changes the position of one or more visible light glass lenses through the lens position adjustment mechanism, thereby changing the virtual image position of the electronic display screen. 9 . 9. The infrared telescope ranging system according to claim 7, further comprising an interface management unit, the interface management unit manages the reception of infrared detector signals, the reception of button module I/O signals and an electronic display screen. data output. 10. An infrared telescope ranging system based on external equipment is characterized in that, comprising external equipment, infrared telescope, and described external equipment comprises training detection model generation module, and described training detection model generation module in the generation detection model process, The target to be measured is classified according to the shape characteristics of the target, and the infrared image of the target to be measured is obtained by shooting the infrared image shooting equipment, and the infrared image data collection is formed from various types of target images. The calibration of the measurement target, the calibration content includes the category of each target to be measured and the rectangular frame that can cover the entire target, establish a detection model, and use the calibrated data as input data to obtain complete model parameters; The infrared telescope includes: infrared Imaging lens, infrared detector, embedded processor module, electronic display screen, eyepiece module, infrared glass lenses with different curvatures and different refractive indices are arranged on the same axis in the infrared telescope to form an infrared imaging lens, and the infrared photosensitive unit is set to be non-infrared. The refrigerated infrared focal plane array forms an infrared detector in an infrared telescope, and the embedded processor module includes a target detection module, a target ranging module, and an image identification module. For the infrared telescope embedded hardware platform used, the training detection model will be trained. After being converted into a format matching the embedded platform, the embedded system of the infrared telescope imports the training detection model. After the infrared telescope obtains the infrared image, it enters the training detection model. The target detection module obtains the target category and coverage in the infrared image. The overall rectangular frame of the target, the target ranging module calculates the target and the detector pixel size according to the preset height of the target to be measured, the pixel height of the target person or object in the screen, the known focal length of the infrared telescope and the size of the detector pixel. The distance between observers, the image identification module obtains the current target distance to form the target distance identification, and the eyepiece module formed by arranging infrared and visible glass lenses with different shapes and different refractive indices, the eyepiece module projects the image on the electronic display screen to the eyepiece module. on the retina of the human eye.

Images