CN102353519B - Resolving power evaluation method for three-generation dim light image intensifier - Google Patents

Abstract

The invention discloses a third-generation low-light image intensifier resolution measuring device and a resolution evaluation method, belonging to the field of optical measurement and measurement. It is characterized in that the resolution measurement device is constructed with light source components, resolution targets, collimator, imaging objective lens, test obscura, CCD camera and computer, and the image intensifier under test images the resolution target irradiated by the standard light source to itself On the fluorescent screen, the frame image of the target line converted by the CCD camera is sent to the computer, and the built-in image processing software of the computer uses the normalized cross-correlation model and the optical modulation model to process the single frame image of the target line successively to obtain a single frame Processing results, and then analyze the multi-frame processing results and perform corresponding supplementary operations to obtain the final resolution evaluation results. The invention solves the objective evaluation problem in the measurement of the resolution of the third-generation low-light image intensifier, can be extended to ICCD testing and other measurement fields that require objective evaluation of the resolution, and has wide application prospects.

Description

Evaluation method for resolution of third-generation low-light image intensifiers

technical field

The invention belongs to the technical field of optical measurement, and mainly relates to a third-generation low-light image intensifier resolution measurement instrument, in particular to a set of third-generation low-light image intensifier resolution measurement device and an evaluation method for the third-generation low-light image intensifier resolution. the

Background technique

In recent years, with the advancement of science and technology, the requirements for low-light technology in the fields of space astronomical detection, fluorescence detection, and military night vision reconnaissance have become higher and higher. The role of the third-generation low-light image intensifier as a core device in this field has become increasingly prominent. The accurate measurement of the parameters of the third-generation low-light image intensifier is of great significance to improve its performance in scientific research and production, and it is also an urgent requirement in many fields of the national defense industry. The resolution is the spatial frequency corresponding to the 2% to 3% modulation degree of the MTF curve of the low-light image intensifier, and it is one of the important parameters reflecting the performance of the low-light image intensifier, which determines the low-light system in the range of 10 ^-3 1x to 10 ^-1 The working distance and image clarity when the illuminance is above 1x. The measurement of the resolution of the low-light image intensifier provides an accurate measurement and calibration method for the development, production and application of the low-light image intensifier.

The traditional method of measuring the resolution of the second-generation low-light image intensifier usually adopts the visual observation method. Nanjing University of Science and Technology, Shijiazhuang Ordnance Institute and the No. 205 Research Institute of China Ordnance Industry have developed the resolution test of the second-generation low-light image intensifier. Equipment, all adopt visual observation method. The second-generation low-light image intensifier resolution test equipment includes standard light source, integrating sphere, neutral attenuation film, power supply, resolution target, collimator, imaging objective lens, low-light image intensifier to be tested, eyepiece, and finally the human eye Observe the resolution target image on the fluorescent screen of the image intensifier to determine the resolution of the image intensifier. This method has the advantages of simple and intuitive measurement, but it is affected by human subjective factors. For example, for the same target, different observers have different observation results. There are some differences; so the measurement accuracy of visual observation method is not high, and the repeatability is poor. Therefore, it is very necessary to objectively evaluate low-light night vision devices. the

In "Optical Technology" May 2000, Vol. 21, No. 2, P451-453, Liu Zhengyun et al. used the grayscale-gradient co-occurrence matrix model to analyze the texture characteristics of the resolution measurement image of the second-generation low-light night vision device. In the analysis, five fringe characteristic parameters such as gray entropy, gradient entropy and gray mean square error are used to assist artificial selection and turning point judgment as the objective criterion of resolution. The calculation accuracy of its algorithm is not high, and the quantitative conclusion cannot be given by the algorithm itself, and subjective judgment must be assisted, and when there are bright or dark spots in the field of view or the size of the fringe image itself is small, it will lead to poor regularity of the results. the

The third-generation low-light image intensifier belongs to the photoelectric imaging device, and the image formed by it has the same problems as all other fluorescent screen imaging, such as screen flicker, dark background noise, ion spots, etc. Therefore, the new performance and indicators of the third-generation low-light image intensifier are very important The resolution measurement puts forward higher requirements, and it is urgent to use the resolution objective evaluation algorithm to develop a low-light image intensifier resolution measurement system to improve the repeatability and accuracy of the three-generation low-light image intensifier resolution measurement. the

At present, there are no public reports on the measurement of the resolution of the third-generation low-light image intensifiers using objective evaluation methods at home and abroad. the

Contents of the invention

The technical problem to be solved by the present invention is to provide a set of devices for measuring the resolution of the third-generation low-light image intensifier and a method for objectively evaluating the resolution of the third-generation low-light image intensifier for the deficiencies of the prior art. the

The third-generation low-light image intensifier resolution measuring device provided by the present invention includes a light source assembly, a resolution target, a collimator, an imaging objective lens with a zoom function, a test obscura with input and output windows, a base platform, and CCD camera and computer with temperature control function, the light source assembly, collimator and test dark box are fixed on the base platform through the corresponding support frame, the resolution target is fixed on the collimator, and the target The target surface is located on the object focal plane of the collimator and the center of the target surface is located on the optical axis of the measurement optical path. The imaging objective lens and the CCD camera are respectively installed on the base platform through a two-dimensional translation mechanism and a three-dimensional translation mechanism. The optical axis of the imaging objective lens and the center of the target surface of the CCD camera are all located on the optical axis of the measurement optical path, and the phosphor screen of the image intensifier to be tested and its center placed in the test dark box correspond to the object focal point of the CCD camera. on the surface and focus; the light beam emitted by the light source assembly illuminates the resolution target, and the target line pattern of the resolution target is collimated by the collimator and then focused by the imaging objective lens to the third-generation low light to be measured The photocathode surface of the image intensifier; the three-generation low-light image intensifier to be tested forms a brighter target line image on its fluorescent screen after multiplying the collimated target line pattern; the CCD camera collects the three-generation low-light image intensifier to be tested The target line image on the fluorescent screen is converted into an electrical signal and transmitted to the computer; the computer controls the CCD camera to complete the setting of relevant test parameters; collects the cold background image, thermal background image and resolution of the CCD camera output The target line image of the target; a series of image processing is performed on the collected corresponding signals, and finally the resolution evaluation results of the three-generation low-light image intensifier to be tested are obtained. the

The three-generation low-light image intensifier resolution evaluation method realized by the three-generation low-light image intensifier resolution measuring device of the present invention comprises the following steps:

The first step, initialize the working parameters of the CCD camera;

In the second step, according to the keyboard instruction, the cold background image _AL and the hot background image A _R output by the CCD camera are successively collected and stored in the memory,

The third step is to collect a frame of resolution target image F output by the CCD camera according to the keyboard instruction and store it in the memory, and display the image on the display screen at the same time;

The fourth step is to call the cold background image _AL , the hot background image A _R and the resolution target image F from the memory according to the keyboard instruction, and subtract the cold background image AL and the hot background image F from the resolution target image _F according to the pixel-by-pixel gray value. The background image A _R corresponds to the grayscale value of the pixel, obtain the image G to be tested and its corresponding two-dimensional grayscale matrix, store the two-dimensional grayscale matrix and display the image G to be tested on the display screen;

The fifth step is to set the image processing area, that is, the ROI area, according to the keyboard command, and the ROI area is the area limited by the highest resolvable or second-highest resolvable resolution target line group of the measured image intensifier;

The sixth step is to generate standard templates corresponding to all target line units in the target line group with the lowest resolution and the target line group with the second-lowest resolution in the ROI area;

The seventh step is to use the generated standard templates to traverse all the pixels in the ROI area in a row-by-row and pixel-by-pixel manner, and at the same time, use the normalized cross-correlation formula to calculate the cross-correlation coefficient between the standard template and the sub-image at each pixel position NC value, one by one to find out the corresponding subgraph with the maximum cross-correlation coefficient NC value with each standard template, that is, the matching target line unit, and record these maximum cross-correlation coefficient NC values and the serial number (i, j) of the corresponding matching target line unit ;Comparing the cross-correlation coefficient NC value of these matching target line units with the resolution threshold NC _resolution , among the matching target line units whose cross-correlation coefficient NC value is greater than the resolution threshold NC _resolution , the matching target line unit with the highest resolution is used as Best matching target line unit;

The eighth step, in the best matching target line unit, take the gray value I ₁ I ₂ ... I _k of a column of pixels in the center of the horizontal target line group and sort these gray value values, and remove k/Q After the maximum value, the selected gray value sequence I _{is selected} and Q is a positive integer. If k/Q is not an integer, the single digit is rounded down, and the modulation degree of the best matching horizontal target line group is calculated with the following formula:

$\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}$

亦即最佳匹配靶线单元的光学调制度；  In the formula, I _avg is the average value of each gray value in the selected gray value sequence, I _max is the average value of each gray value greater than the average value in the selected gray value sequence, and I _min is the average value of the selected gray value sequence that is smaller than the average value The average value of each gray value; similarly, calculate the modulation degree of the best matching vertical target line group; find the average value of the modulation degree of the best matching target line unit horizontal target line group and vertical target line group

That is, the optical modulation degree that best matches the target line unit;

与调制度阈值M_c进行比较：若 

继续判断最佳匹配靶线单元的光学调制度 

是否明显高于阈值M_c，若 

则最佳匹配靶线单元对应的分辨力值为单帧处理最终结果；如果 

则计算分辨力高于最佳匹配靶线单元的相邻单元的光学调制度 如果高分辨力相邻靶线单元的光学调制度 

满足 

则高分辨力相邻靶线单元对应的分辨力值为单帧处理的最终结果，否则维持最佳匹配靶线单元对应的分辨力值为单帧处理最终结果；若 

计算分辨力低于最佳匹配靶线单元的相邻单元的光学调制度 

如果满足 且 

则低分辨力相邻靶线单元对应的分辨力值为单帧处理最终结果；如果满足 

则再次利用归一化互相关公式计算低分辨力相邻靶线单元及其对应标准模板之间的互相关 系数NC_相邻，如果NC_相邻≥0.97NC，则最佳匹配靶线单元对应的分辨力值为单帧处理最终结果，如果NC_相邻＜0.97NC，则低分辨力相邻靶线单元对应的分辨力值为单帧处理最终结果并在显示屏上显示；  The ninth step is to best match the optical modulation of the target line unit

Compare with the modulation degree threshold value M _c : if

Continue to judge the optical modulation degree of the best matching target line unit

Is it significantly higher than the threshold M _c , if

Then the resolution value corresponding to the best matching target line unit is the final result of single frame processing; if

Then calculate the optical modulation degree of the adjacent unit whose resolution is higher than the best matching target line unit If the optical modulation degree of high-resolution adjacent target line unit

satisfy

Then the resolution value corresponding to the high-resolution adjacent target line unit is the final result of single-frame processing, otherwise the resolution value corresponding to the best matching target line unit is the final result of single-frame processing; if

Calculate the optical modulation of adjacent elements with lower resolution than the best matching target line element

if satisfied and

Then the resolution value corresponding to the low-resolution adjacent target line unit is the final result of single-frame processing; if it satisfies

Then use the normalized cross-correlation formula to _calculate the _cross -correlation coefficient between the low-resolution adjacent target line unit and its corresponding standard template. The resolution value is the final result of single-frame processing. If the NC _{is adjacent to} <0.97NC, the resolution value corresponding to the low-resolution adjacent target line unit is single-frame processing and the final result is displayed on the display;

并逐一与调制度判据Mc比较，光学调制度 

最接近M_c的靶线单元所对应的分辨力值视为所述三代微光像增强器分辨力评价方法的最终结果并在显示屏上显示。  In the tenth step, repeat the third step to the ninth step several times to process the subsequent multi-frame resolution target images respectively, so as to obtain a corresponding number of single-frame resolution processing results, if there are half or more of the single-frame resolution processing results Same, take this result as the final result of the evaluation method for the resolution of the third-generation low-light image intensifier; if the single-frame resolution processing results of multi-frame images are discretely distributed, the target line unit corresponding to each single-frame resolution processing result Calculate the degree of optical modulation according to the eighth step

And compare with the modulation degree criterion Mc one by one, the optical modulation degree

The resolution value corresponding to the target line unit closest to _Mc is regarded as the final result of the evaluation method for the resolution of the third-generation low-light image intensifier and displayed on the display screen.

The overall technical effects of the present invention are reflected in the following aspects. the

(1) The present invention establishes a set of three-generation low-light image intensifier resolution measuring device through the light source assembly, resolution target, collimator, imaging objective lens, CCD camera and computer, wherein, the light source assembly adopts uniform diffuse weak illumination The light source solves the problem of poor light source uniformity in the background of low-light measurement in the prior art, and lays the foundation for objectively evaluating and measuring the resolution of the third-generation low-light image intensifier; in addition, the CCD camera used has a refrigeration unit and a temperature controller, It greatly eliminates the electronic noise and thermal noise of the CCD itself, and provides hardware guarantee for software analysis and processing of noise like the third-generation low-light intensifier. the

(2) In the present invention, image processing software is installed in the computer, and the image processing software objectively evaluates the resolution target images formed by three generations of low-light intensifiers and CCD cameras successively, and its evaluation algorithm adopts a method based on different principles The dual model, that is, the normalized cross-correlation model and the optical modulation model, the two models are independent and complementary to each other, and the final conclusion is obtained through double independent criteria. Therefore, the invention can accurately measure the resolution of the third-generation low-light image intensifier, has high measurement accuracy and good repeatability, and provides reliable technical support for the development, production and application of the third-generation low-light image intensifier. the

(3) In the process of image processing, the matching coefficient in the normalized cross-correlation model is used as an independent criterion, which can not only qualitatively measure whether the reticle group on the target surface is clear, but also search for the image corresponding to a certain resolution. The position of the target line group is used to lock the image processing area for subsequent calculations. As a result, not only the computational load of image processing is reduced, but also erroneous results caused by unnecessary calculations are avoided. the

(4) The present invention introduces the optical modulation degree model into the objective evaluation of resolution in an original way, because under low-contrast conditions in the low-light field, the image resolution determined by the optical modulation degree criterion is consistent with the result of human eye observation It has a high degree of consistency. Therefore, taking the degree of optical modulation as another independent criterion of the present invention, the resolution data can be obtained quantitatively, thereby improving the objectivity, precision and accuracy of the evaluation method of the present invention. the

Description of drawings

Fig. 1 is a schematic diagram of the composition of the third-generation low-light image intensifier resolution measuring device of the present invention. the

Fig. 2 is a measurement light path diagram of the third-generation micro-light image intensifier resolution measurement device. the

Fig. 3 is a working flow chart of the method for evaluating the resolution of the third-generation low-light image intensifier of the present invention. the

Detailed ways

The present invention will be described in further detail below in conjunction with the accompanying drawings and preferred examples. the

As shown in Figure 1, the preferred example of the third-generation low-light image intensifier resolution measurement device mainly includes a light source assembly 1, a resolution target 2, a collimator 3, an imaging objective lens 4, a test obscura 5, a CCD camera 6, and a base platform 7. Computer 8 and power supply 9. the

The light source assembly 1 is composed of a standard light source 1-1, a neutral filter, an iris diaphragm 1-2 and an integrating sphere 1-3. The standard light source 1-1 selects the halogen tungsten lamp, the neutral filter and the variable diaphragm attenuate the light beam emitted by the halogen tungsten lamp, and after being integrated by the integrating sphere 1-3, a light beam with a color temperature of 2856K is formed at the exit. The range of illuminance is between 10 ^-3 1x and 10 ^-1 1x, and the unevenness in the effective illumination area is less than 1%. In the measurement of low-light image intensifier resolution, due to the high sensitivity of the human eye, the uniformity of the low-light source has little influence on the measurement results, and the sensitivity of the CCD camera is lower than that of the human eye, so the uniformity of the low-light source is required very high. In the traditional process, the uniformity of the low-illuminance light source is one of the key indicators that is difficult to guarantee. Due to the uneven illumination of the light source during the measurement, each area of the target surface presents an uneven illuminance, which leads to a decrease in the contrast of the image intensifier, thereby introducing inestimable Uncertain factors also have unpredictable effects on the experimental results. Therefore, in the present invention, the power source 9 for supplying power to the standard light source 1-1 selects a high-precision, high-stability constant current and constant voltage source to meet the requirements of light source stability and color temperature change. Integrating sphere 1-3 is the key to forming a diffuse reflection light source, and it consists of two hemispherical shells whose inner walls are coated with a white diffuse reflection layer. The relevant parameters of the integrating sphere 1-3 in the present invention are determined by the calculation formula of illuminance diffusion, and its diameter is taken as 260mm, and the diameter of the light exit hole is 20mm and the area is 7.7% of the total area of the integrating sphere, which should not exceed the total area of the integrating sphere. 10% required.

的比例缩小。  The resolution target 2 is the USAF1951 resolution target frequently used in the field of low-light testing. light and dark stripes. Each group of target line units is composed of three horizontal target lines and three vertical target lines of equal length, the length of the target line is five times the width of the target line, and the width of the target line is equal to the interval between adjacent target lines. The distance between the horizontal target line and the vertical target line is twice the width of the target line. Starting from the largest group of target line units, every two groups of target line units with similar sizes are divided by

ratio is reduced.

The collimator 3 is a F1000 zoom lens with a double-separated objective lens with excellent image quality inside. Its focal length is 1000mm, the effective aperture is Φ100mm, the resolution is 1.3″, the field of view is 1°38′, and the parallax is ≤0.20mm. Imaging objective lens 4 has a zoom function, its focal length is 100mm, and the effective aperture is Φ80mm. There are light-transmitting windows on the front and rear side walls of the test dark box 5, and the tested image intensifier 5-1 is supported in the cavity of the test dark box 5 by corresponding fixtures. inside.

The light source assembly 1, the collimator 3 and the test obscura 5 are fixed on the base platform 7 through corresponding support frames 7-1, 7-2, 7-4, and the resolution target 2 is fixedly connected to the collimator 3, and its The target plane is located on the focal plane of the object side of the collimator 3 and the center of the target plane is located on the optical axis of the collimator 3, that is, the optical axis of the measurement optical path. Also located on the optical axis of the measuring beam path. The imaging objective lens 4 and the CCD camera 6 are respectively installed on the base platform 7 through the two-dimensional translation mechanism 7-3 and the three-dimensional translation mechanism 7-5. During the test, by adjusting the two-dimensional translation mechanism 7-3 and the three-dimensional translation mechanism 7-5, The optical axis of the imaging objective lens 4 and the center of the target surface of the CCD camera 6 are located on the optical axis of the measurement optical path, and the fluorescent screen of the image intensifier 5-1 to be measured is located on the object focal plane of the CCD camera 6. the

As shown in Figure 2, the light beam emitted by the standard light source 1-1 passes through the neutral filter, the iris 1-2 and the integrating sphere 1-3 successively to form a uniform diffused light beam and illuminate the resolution target 2; All the target line units on the resolution target 2 are collimated by the collimator 3 and focused by the imaging objective lens 4 onto the photocathode surface of the third-generation low-light image intensifier 5-1 to be tested in the test dark box 5; The optical image intensifier 5-1 multiplies the collimated target line pattern to form a brighter target line image on its fluorescent screen; the CCD camera 6 collects the target line image on the fluorescent screen and converts it into an electrical signal and transmits it to the computer 8 . the

The CCD camera 6 is a COOLSNAP K4 camera produced by American Photometric Company and is controlled by a computer 8. The camera has a refrigeration unit and a temperature controller, and the highest sampling frequency is 20MHz; the pixel size is 7.4μm×7.4μm, and the number of pixels is 2048×2048. The chip area is 15.16m×15.16mm, and the frame rate is 3fps. When the CCD camera 6 is in working order, the image collected when its shutter is closed is a cold background image, which is actually the inherent thermal noise and electronic noise of the CCD camera 6 itself; when its shutter is opened and the standard light source 1- 1 When the power is not turned on, the collected image is the thermal background image, which is generated by the test environment; when the shutter is opened and the standard light source 1-1 is turned on, the collected image is the target line image. The cold background image, hot background image and target line image are all in bitmap format. the

The computer 8 has a built-in acquisition card, memory and image processing software and is equipped with a mouse and a keyboard, and the computer 8 is connected with the CCD camera 6 through a PCI bus. The initial working state data of the CCD camera and the acquisition card, the modulation degree threshold M _c , the resolution target line unit and the resolution value correspondence table are prestored in the memory, and various images output by the CCD camera 6 can also be stored. The function of the image processing software is to control the CCD camera 6 to complete the setting of relevant test parameters; collect the cold background image output by the CCD camera 6, the thermal background image and the target line image of the resolution target 2; carry out a series of corresponding images collected Image processing, and finally obtain an objective evaluation result of the resolution of the three-generation low-light image intensifier 5-1 to be tested.

The resolution evaluation method of the present invention is implemented by the computer 8. After the third-generation low-light image intensifier 5-1 to be tested is placed in the resolution measurement device, the image processing software executes the following operation steps according to the flow shown in FIG. 3 .

The first step, initialization

After power on, load the drivers of CCD camera, capture card, mouse and keyboard, and initialize the device. In conjunction with the input data of the keyboard, the working state of the CCD camera 6 and the capture card is set, as no keyboard input, then call the CCD camera in the storage memory and the initial working state data of the capture card, and set the working state of the two; this default working state is In focus mode and video monitoring mode, the lightness and contrast parameters are both 50, the working temperature is -25°C, the acquisition frequency is 3 frames per second, the acquisition mode is single frame acquisition, and the transmission rate is 20M/s;

The second step is to collect cold background images and hot background images

After receiving the acquisition order of keyboard input, at first send cold background acquisition command to CCD camera 6, gather the cold background image _AL of CCD camera 6 output by PCI bus and store in memory; Then, send thermal background to CCD camera 6 The collection command, when the light source 1-1 is not turned on, collects the thermal background image _AR output by the CCD camera 6 through the PCI bus and stores it in the memory.

The third step is to collect the resolution target image

After receiving the acquisition command input by the keyboard, send the resolution target acquisition command to the CCD camera 6, and when the light source 1-1 is turned on, collect a frame of resolution target image F output by the CCD camera 6 through the PCI bus and store it in In the memory, the resolution target image F is displayed on the display screen at the same time. the

The fourth step, image preprocessing

In the present invention, image preprocessing refers to a processing method in which the resolution target image F is subtracted from the pixel gray value corresponding to the cold background image _AL and the hot background image A _R according to the pixel gray value. The specific algorithm formula as follows:

g(s, t)=n(s, t)-n _L (s, t)-n _R (s, t) where, n(s, t), n _L (s, t), n _R ( s, t) are the gray values of the resolution target image F, the cold background image _AL , and the hot background image A _R at the pixel point (s, t), and g(s, t) is the measured value obtained after preprocessing The gray value of the image G at the pixel point (s, t). After the preprocessing of the entire frame of images is completed, the two-dimensional grayscale matrix of the image G to be tested is obtained and stored in the memory, and the image G to be tested is displayed on the display screen at the same time.

The fifth step is to set the ROI area

After receiving the ROI area setting command input by the keyboard, the size of the ROI area is calculated according to the corresponding formula. The ROI area is the working area, which refers to the image area that needs to be processed. This area is a square area defined by the highest resolution target group of the measured image intensifier. Since the resolution target in this embodiment is the USAF1951 target, the target includes 10 groups of -2, -1, 0, 1...6, 7, and the side length L _i of the ROI area corresponding to each group Determined by:

${\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; [</span>\frac{\mathrm{<span\; class="notranslate">\; 7</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\sqrt[\mathrm{<span\; class="notranslate">\; 6</span>}]{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\sqrt[\mathrm{<span\; class="notranslate">\; 6</span>}]{\mathrm{<span\; class="notranslate">\; 32</span>}}<span\; class="notranslate">\; ]</span>$

In the formula, i represents the group of resolution targets, that is, i=-2, -1, 0, 1...6,7. At present, the highest resolution of the third-generation image intensifiers is generally in the second group. Therefore, in this embodiment, corresponding to i=2, L ₂ =3.56 is obtained through calculation. Defining the ROI area can avoid dealing with the low-resolution target line area and reduce the amount of calculation.

The sixth step is to generate the target line unit standard template

The standard template refers to the ideal target line image generated according to the target line number, shape and size of each discrete target line unit on the used resolution target. the

For this embodiment, the width of the standard template is determined by the following formula:

${\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 12</span>}{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\sqrt[\mathrm{<span\; class="notranslate">\; 6</span>}]{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; j</span>}}$

In the formula, j represents the serial number of each target line unit in each set of resolution target lines. In this embodiment, the target line group with the lowest resolution in the ROI area is the second group, that is, i=2, the value range of j is 1, 2, 3...6, and j=1 corresponds to the resolution The target line unit with the largest line width in the force group, j=6 corresponds to the target line unit with the smallest line width in the resolution force group. For the USAF1951 resolution target, the height of the standard template is 5/12 of the width, there are three horizontal and three vertical rectangular target lines in the template, the six target lines are equal in size, the aspect ratio of the target lines is 5:1, and the width of the target lines is E _{i, j} (unit mm) is determined by the following formula:

${\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\sqrt[\mathrm{<span\; class="notranslate">\; 6</span>}]{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; j</span>}}$

E _{i, j} is also the spacing between three horizontal target lines or three vertical target lines distributed equidistantly, and the distance between the horizontal target line group and the vertical target line group is 2E _{i, j} .

According to the above calculation formula and the corresponding target line parameter relationship, generate standard templates corresponding to all target line units in the target line group with the lowest resolution and the target line group with the second-lowest resolution in the ROI area. For this embodiment, the standard templates corresponding to the units of i=2 and j=1-6, i=3 and j=1-6 are sequentially generated. the

The seventh step is to calculate the normalized cross-correlation coefficient

Place the generated standard template in the upper left corner of the ROI area of the image G to be tested, and move the standard template row by row and pixel by pixel, and use the normalized cross-correlation formula to calculate the normalized correlation between the current standard template and the sub-image. Normalize the cross-correlation coefficient (referred to as the cross-correlation coefficient) NC until the current standard template traverses all the pixels in the ROI area. The sub-image is the part of the image G to be tested covered by the standard template. Then find out the subgraph corresponding to the maximum cross-correlation coefficient NC value, and record the maximum cross-correlation coefficient NC value and its corresponding subgraph position. The specific method is to use the standard templates to traverse all the pixels in the ROI area one by one in accordance with the order of the standard templates from low to high resolution, and finally find the corresponding subgraph with the largest cross-correlation coefficient NC value with each standard template, that is, the matching target Line unit, record these maximum cross-correlation coefficient NC values and the target line unit numbers (i, j) corresponding to the corresponding subgraphs. the

The larger the NC value of the cross-correlation coefficient, the more similar the subgraph is to the standard template. When the NC value of the cross-correlation coefficient is 1, it is the ideal matching position. In actual situations, the cross-correlation coefficient NC cannot be 1, so only If the subimage position corresponding to the maximum cross-correlation coefficient NC value is found on the image G to be tested, the target line unit at this position can be considered as the matching target line unit of the current standard template. In this way, each standard template corresponds to a matching target line unit. Compare the cross-correlation coefficient NC value of these matching target line units with the resolution threshold NC _resolution , and find the target line unit with the highest resolution among the matching target line units whose cross-correlation coefficient NC value is greater than the resolution threshold NC _resolution and use it as Best match target line unit.

Among the matching target line units corresponding to all standard templates, if the cross-correlation coefficient NC value is too small, the target line unit is actually indistinguishable, and when the cross-correlation coefficient NC value is too large, it cannot truly reflect the measured Therefore, it is necessary to set an appropriate cross-correlation coefficient NC value, that is, the resolution threshold NC _resolution , to find target line units that are both resolvable and high in resolution. Usually, it is necessary to obtain a resolvable and high-resolution target line unit through a large number of experiments, and then use the cross-correlation coefficient NC value corresponding to the unit as the resolution threshold NC _{for resolution} . The NC _resolution generally ranges from 0.5 to 0.7, and the NC _resolution is 0.68 in this embodiment.

Since the low-light image intensifier under test uses fluorescent screen imaging, there are inherent bright and dark flickering and screen jittering flickering phenomena. For this reason, the present invention uses the characteristics of high accuracy, good adaptability, and insensitivity to the linear transformation of the image gray value of the normalized cross-correlation model to perform matching operations on the standard template and the subgraph of the image G to be tested covered by it , so as to reduce the impact of bright and dark flicker on the screen of the image intensifier under test on the evaluation of resolution measurement; in addition, since the calculation result of the normalized cross-correlation model is to solve the best matching target line unit, it can be well Solve the impact of screen jitter and flicker on the evaluation of resolution measurement. the

The eighth step, calculate the degree of optical modulation

The size of the optical modulation reflects the resolvability of the image to a large extent, and the clarity of the image can be judged as a whole. The sharpness of different target lines is directly proportional to the optical modulation degree of its gray scale curve. At the same time, there is a good consistency between the degree of optical modulation and the clarity of the image observed by the human eye. The higher the degree of optical modulation, the clearer the image, and vice versa. In the present invention, the degree of optical modulation is calculated according to the following steps:

8.1 In the best matching target line unit, take the gray value I ₁ I ₂ ... I _k of a column of pixels at the center of the horizontal target line group and sort these gray value values, and remove k/Q maximum values Afterwards, the selected gray value sequence I _{is obtained} and Q is a positive integer. If k/Q is not an integer, the single digit is rounded down. The Q value depends on the number of ion spots on the fluorescent screen of the image intensifier under test. The size of the Q value should not only ensure the effective removal of ion spots, but also ensure that as many pixels as possible are left for subsequent processing. In this embodiment, Q=12.

_Calculate the arithmetic mean value for the selected gray value sequence I to obtain I _avg ; _select the gray value greater than the average value I _avg for the selected gray value sequence I to calculate the arithmetic mean value to obtain I _max , and calculate the gray value for the gray value smaller than I _avg The arithmetic mean yields I _min . After that, use the following formula to calculate the modulation degree of the horizontal target line group:

$\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}$

8.2 Take the gray value of a row of pixels at the center of the vertical target line group of the best matching target line unit, and use the method in 8.1 to calculate the modulation degree of the vertical target line group. the

即最佳匹配靶线单元的光学调制度。将该靶线单元的光学调制度 

与存储器中的调制度阈值M_c进行比较，若 

表明该最佳匹配靶线单元在调制度判据下可分辨，则进入9.1步骤；若 表明该最佳匹配靶线单元在调制度判据下不可分辨，则进入9.2步骤。  8.3 Calculate the mean value of the modulation degree of the horizontal target line group and the vertical target line group of the best matching target line unit

That is, the optical modulation degree that best matches the target line unit. The optical modulation degree of the target line unit

Compared with the modulation degree threshold M _c in the memory, if

It shows that the best matching target line unit can be distinguished under the modulation degree criterion, then go to step 9.1; if If it shows that the best matching target line unit is indistinguishable under the modulation degree criterion, then go to step 9.2.

The modulation degree threshold M _c is the critical point of discrimination determined after repeated experiments and repeated verifications. Generally speaking, if the modulation degree is lower than 0.1, the resolution target image will be difficult to distinguish. Therefore, the value of the modulation degree threshold M _c is generally in the range of 0.1≤M _c ≤0.2.

Since the third-generation low-light image intensifiers are fluorescent screen imaging, there are inherent ion spots. For this reason, in the calculation of the optical modulation degree, the present invention can effectively eliminate the large abnormal gray value caused by the ion spot by removing the k/Q maximum gray value, thereby removing the interference of the ion spot on the measurement and evaluation of the resolution . the

The ninth step is to obtain the result of calculating the resolution of a single frame image

是否明显高于阈值M_c，若 

则最佳匹配靶线单元的序号为单帧处理结果；如果 

则计算分辨力高于最佳匹配靶线单元的相邻单元的光学调制度 

如果高分辨力相邻靶线单元的光学调制度 满足 

则该高分辨力靶线单元的序号为单帧处理结果，否则维持最佳匹配靶线单元的序号为单帧处理结果。  9.1 When Continue to judge the optical modulation degree of the best matching target line unit

Is it significantly higher than the threshold M _c , if

Then the sequence number of the best matching target line unit is the single frame processing result; if

Then calculate the optical modulation degree of the adjacent unit whose resolution is higher than the best matching target line unit

If the optical modulation degree of high-resolution adjacent target line unit satisfy

Then the serial number of the high-resolution target line unit is the single frame processing result, otherwise the serial number of the best matching target line unit is maintained as the single frame processing result.

计算分辨力低于最佳匹配靶线单元的相邻单元的光学调制度 

如果满足 

且 

则该低分辨力相邻靶线单元的序号为单帧处理结果；如果满足 

则进一步利用归一化互相关公式计算低分辨力相邻靶线单元及其对应标准模板之间的互相关系数NC_相邻，如果互相关系数NC_相邻和最佳匹配靶线单元的互相关系数NC比较也无较大变化，即NC_相邻≥0.97NC，则最佳匹配靶线单元的序号为单帧处理结果，如果NC_相邻＜0.97NC，则低分辨力相邻靶线单元的序号为单帧处理结果。  9.2 When

Calculate the optical modulation of adjacent elements with lower resolution than the best matching target line element

if satisfied

and

Then the serial number of the low-resolution adjacent target line unit is the single-frame processing result; if it satisfies

Then the normalized cross-correlation formula is further used to calculate the cross-correlation coefficient between the low-resolution adjacent target line unit and its corresponding standard template NC _{is adjacent} , if the cross-correlation coefficient NC _{is adjacent to} and best matches the cross-correlation of the target line unit There is no big change in the comparison of the number of NCs, that is, if the NC _{is adjacent to} ≥0.97NC, the serial number of the best matching target line unit is the result of single-frame processing; if the NC _{is adjacent to} <0.97NC, the number of the adjacent target line unit with low resolution The serial number is the single frame processing result.

The serial number of the above-mentioned resolution target line unit is expressed as the jth unit of the i-th group, and the resolution value corresponding to the serial number needs to be found by looking up the table (Table 1), and the resolution value is taken as the final processing result of a single frame and displayed on the screen. the

The tenth step is to obtain the result of multi-frame image resolution calculation

In order to make the calculation result more accurate, multi-frame acquisition is performed on a resolution target image, usually 10 frames are selected, and then steps 3 to 9 are repeated multiple times (10 times) to perform processing, so as to obtain a corresponding number of single-frame resolution processing results. In this way, there will be the following two situations:

1). If half or more of the single-frame resolutions have the same final processing result, the result can be used as the final result of the evaluation method for the resolution of the third-generation low-light image intensifier. the

并逐一与调制度判据M_c比较，光学调制度 

最接近M_c的靶线单元对应的分辨力值视为三代微光像增强器分辨力评价方法的最终结果。  2). The final processing results of the single-frame resolution of multi-frame images are discretely distributed, that is, the proportion of the final processing results of no single-frame resolution exceeds half. At this time, the target line unit corresponding to each single-frame resolution processing result Calculate the degree of optical modulation according to the eighth step

And compared with the modulation degree criterion M _c one by one, the optical modulation degree

The resolution value corresponding to the target line unit closest to _Mc is regarded as the final result of the evaluation method of the resolution of the third-generation low-light image intensifier.

Table 1 Correspondence between resolution target line unit and resolution value

Unit: Line Pair

Claims (2)

1. A three-generation low-light image intensifier resolution evaluation method, the method is realized on three-generation low-light image intensifier resolution measuring devices, and the three-generation low-light image intensifier resolution measurement devices include light source components (1 ), resolution target (2), collimator (3), imaging objective with zoom function (4), test obscura with input and exit windows (5), base platform (7), temperature CCD camera (6), computer (8) with control function, the light source assembly (1), collimator (3) and test obscura (5) through the corresponding support frame (7-1, 7-2, 7- 4) fixed on the base platform (7), the resolution target (2) is fixed on the collimator (3), its target surface is located on the object focal plane of the collimator (3) and The center of the target surface is located on the optical axis of the measurement optical path, and the imaging objective lens (4) and the CCD camera (6) are respectively installed on the The base platform (7), the optical axis of the imaging objective lens (4) and the center of the target surface of the CCD camera (6) are all located on the optical axis of the measurement optical path, and the image to be measured in the described test dark box (5) is intensified The phosphor screen of the device (5-1) and its center are correspondingly located on the object focal plane and the focal point of the CCD camera (6); the light beam sent by the light source assembly (1) illuminates the resolution target (2), The target line pattern of the resolution target (2) is collimated by the collimator (3) and then focused by the imaging objective lens (4) to the photocathode of the third-generation low-light image intensifier (5-1) to be tested On the surface; the three-generation low-light image intensifier (5-1) to be tested forms a brighter target line image on its fluorescent screen after multiplying the collimated target line pattern; the CCD camera (6) collects the three-generation low-light image to be measured The target line image on the fluorescent screen of the intensifier (5-1) and convert it into an electrical signal; It is characterized in that: the computer (8) evaluates the resolution of the three-generation low-light image intensifier according to the following steps: The first step, initialize the operating parameters of the CCD camera (6); In the second step, according to the keyboard instruction, the cold background image _AL and the thermal background image A _R of the output of the CCD camera (6) are collected successively and are all stored in the memory; The 3rd step, according to keyboard instruction, gathers a frame resolution target image F that described CCD camera (6) outputs and deposits in memory, shows this image on display screen simultaneously; The fourth step is to call the cold background image _AL , the hot background image A _R and the resolution target image F from the memory according to the keyboard instruction, and subtract the cold background image AL and the hot background image F from the resolution target image _F according to the pixel-by-pixel gray value. The background image A _R corresponds to the grayscale value of the pixel, obtain the image G to be tested and its corresponding two-dimensional grayscale matrix, store the two-dimensional grayscale matrix and display the image G to be tested on the display screen; The fifth step is to set the image processing area, that is, the ROI area, according to the keyboard command, and the ROI area is the area defined by the highest resolvable or second-highest resolvable resolution target line group of the measured image intensifier; The sixth step is to generate standard templates corresponding to all target line units in the target line group with the lowest resolution and the target line group with the second-lowest resolution in the ROI area; The seventh step is to use the generated standard templates to traverse all the pixels in the ROI area in a row-by-row and pixel-by-pixel manner, and at the same time, use the normalized cross-correlation formula to calculate the cross-correlation coefficient between the standard template and the sub-image at each pixel position NC value, one by one to find out the corresponding subgraph with the maximum cross-correlation coefficient NC value with each standard template, that is, the matching target line unit, and record these maximum cross-correlation coefficient NC values and the serial number (i, j) of the corresponding matching target line unit ;Comparing the cross-correlation coefficient NC value of these matching target line units with the resolution threshold NC _resolution , among the matching target line units whose cross-correlation coefficient NC value is greater than the resolution threshold NC _resolution , the matching target line unit with the highest resolution is used as Best matching target line unit; The eighth step, in the best matching target line unit, take the gray value I ₁ I ₂ ... I _k of a column of pixels in the center of the horizontal target line group and sort these gray value values, and remove k/Q After the maximum value, the selected gray value sequence I _{is selected} and Q is a positive integer. If k/Q is not an integer, the single digit is rounded down, and the modulation degree of the best matching horizontal target line group is calculated with the following formula: $\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}$ In the formula, _Imax is the average value of each grayscale value in the selected grayscale value sequence that is greater than the mean value of the selected grayscale value sequence, and _Imin is the average value of each grayscale value in the selected grayscale value sequence that is smaller than the average value of the selected grayscale value sequence Average value; similarly, calculate the modulation degree of the best matching vertical target line group; find the mean value of the modulation degree of the best matching target line unit horizontal target line group and vertical target line group

That is, the optical modulation degree that best matches the target line unit; The ninth step is to best match the optical modulation of the target line unit

Compare with the modulation degree threshold value M _c : if

Continue to judge the optical modulation degree of the best matching target line unit

Is it significantly higher than the threshold M _c , if

Then the resolution value corresponding to the best matching target line unit is the final result of single frame processing; if

Then calculate the optical modulation degree of the adjacent unit whose resolution is higher than the best matching target line unit

If the optical modulation degree of high-resolution adjacent target line unit

satisfy

Then the resolution value corresponding to the high-resolution adjacent target line unit is the final result of single-frame processing, otherwise the resolution value corresponding to the best matching target line unit is the final result of single-frame processing; if

Calculate the optical modulation of adjacent elements with lower resolution than the best matching target line element

if satisfied

and

Then the resolution value corresponding to the low-resolution adjacent target line unit is the final result of single-frame processing; if it satisfies

Then use the normalized cross-correlation formula to _{calculate the} cross-correlation coefficient between the low-resolution adjacent target line unit and its corresponding standard template. The resolution value is the final result of single-frame processing. If the NC _{is adjacent to} <0.97NC, the resolution value corresponding to the low-resolution adjacent target line unit is single-frame processing and the final result is displayed on the display; In the tenth step, repeat the third step to the ninth step several times to process the subsequent multi-frame resolution target images respectively, so as to obtain a corresponding number of single-frame resolution processing results, if there are half or more of the single-frame resolution processing results Same, take this result as the final result of the evaluation method for the resolution of the third-generation low-light image intensifier; if the single-frame resolution processing results of multi-frame images are discretely distributed, the target line unit corresponding to each single-frame resolution processing result Calculate the degree of optical modulation according to the eighth step

And compared with the modulation degree criterion M _c one by one, the optical modulation degree

The resolution value corresponding to the target line unit closest to _Mc is regarded as the final result of the evaluation method for the resolution of the third-generation low-light image intensifier and displayed on the display screen. 2. The three-generation low-light image intensifier resolution evaluation method according to claim 1, characterized in that: the value of the resolution _threshold NC is within the range of 0.5 to 0.7; the value of the modulation degree threshold _Mc In the range of 0.1≤M _c ≤0.2.

Images