Research on Automatic Test of Imaging Resolution of Transmission CCD Camera

Abstract: The automatic test method of CCD camera imaging resolution is researched. In the test plan, the camera resolution target is improved and the data analysis and processing are carried out. The automatic test of the resolution of the CCD camera is realized through the combination of software and hardware. The results The data has been tested and verified to be valid. The experimental results show that the method is objective and accurate, can avoid the influence of subjective factors caused by the previous manual test resolution, and has high practicability.

  0 Preface

At present, the transmission type CCD camera has replaced the traditional film camera as the mainstream photography equipment. However, the measurement of the camera's imaging resolution index by the manufacturers basically adopts the test method based on manual interpretation.

Manual interpretation of the test resolution is simple and convenient for film cameras. However, due to the different visual sensitivity of different human eyes and the difference in detection conditions, it is inevitable to introduce different degrees of subjective error, and it is often difficult to achieve a unified measurement result, which affects Test accuracy.

For a CCD camera, the digital image generated by it for a specific target can be used to implement efficient data analysis and processing technology to automatically achieve a quantitative test of the camera resolution, so as to objectively determine the image quality of the camera.

1 Theoretical analysis

The factors that affect the imaging resolution of CCD cameras mainly include: optical systems, CCD devices, and corresponding circuit processing systems. Among them, the optical system can use interference detection method or transfer function to test its image quality, so as to objectively obtain the corresponding resolution quantification result; the theoretical limit resolution of the CCD device itself can be directly calculated according to its pixel size; Circuit processing system, under ideal circumstances, its impact on image resolution testing is negligible, so it will not be considered here. Based on the above factors, the ideal resolution N of the CCD camera can be calculated by the following formula:



where: N light is the resolution of the optical system; NCCD is the resolution of the CCD device.

Although the above calculations can estimate the theoretical resolution of the entire CCD camera, it is often difficult to accurately reflect the final imaging level of the camera due to assembly errors, system control errors, and subjective uncertainty caused by manual interpretation tests. It is necessary to implement a quantitative test of the resolution index during the inspection of the CCD camera, so as to objectively and comprehensively reflect the imaging quality of the CCD camera.

For this reason, this paper proposes an automatic resolution test method based on the contrast analysis of the raster target image. The method is to use the CCD camera as a light energy information transmission system. According to the test principle of the system transfer function and the definition of sine series expansion, the rectangular distribution function is expanded into a superposition of sinusoidal distributions of different frequencies. Then the contrast transfer function can be expressed as:



Since the photodetector converts the luminous flux into an electrical signal, all high-order harmonic components can be filtered out using electronic methods, so the obtained transfer function relationship becomes:



M and M0 are output contrast (also called modulation) and input contrast, respectively. Suppose the target image light intensity is Imax and the light intensity is Imin, then the contrast is defined as:



In the actual imaging process of the CCD camera, in addition to the target image, there will be light intensity Ib generated by the background brightness. At this time, the definition of contrast is corrected as:



Under laboratory conditions, the target used is high-contrast, and the corresponding M0 =1 can be defined. In this way, the principle of measuring the resolution of the CCD camera is transformed into the contrast analysis and calculation of the resolution target grating pattern image. The confirmed resolvable benchmark thresholds are compared to automatically determine the resolution level reached by the camera, and the resolution quantification result can be quickly obtained by system conversion.

2 Test system composition and test process

The composition of the CCD camera resolution test system is shown in Figure 1. It mainly includes a standard light source, grating target, collimator, image capture card, image memory, industrial computer, and data analysis and processing software.



The initial position of each device must be confirmed before the system test. The target must be accurately calibrated at the focal plane of the collimator. The exit of the collimator corresponds to the entrance of the camera. The optical axis of the collimator and the optical axis of the camera optical system must be the same. At the same time, ensure that the CCD pixel is aligned with the center line of the target stripe.

During the test, turn on the integrating sphere to evenly illuminate the rectangular grating target, start the camera to take pictures, the target is imaged on the focal plane CCD through the collimator and the camera optical system, the light signal is converted into electrical signal, and then the image capture card is used in real time The image data stream generated by the collection is stored in the image memory in the corresponding format under the control of the industrial computer. Image data comprehensive analysis and processing software can directly display data format images, and can also zoom in or zoom out the entire image or selected partial areas. More importantly, it can realize the resolution discrimination calculation of the randomly selected image target image area.

For automatic resolution testing, first determine the distinguishable contrast benchmark. At present, the commonly used resolution criteria mainly include the Rayleigh criterion and the Daws criterion. According to the Rayleigh criterion, when the contrast of the light intensity between the two diffraction spots is less than 15%, the human eye cannot distinguish between the two points; According to the Dawes criterion, when the contrast of the light intensity between the two diffraction spots is 1.55%, it is the resolution limit of the human eye. In order to reasonably determine the test criteria, this article has collected a large number of photographic test images of different models of cameras under different conditions, including dozens of image data that have reached a certain level of resolution after interpretation and review. The contrast calculation and analysis of the acquired image data show that the conditions for determining the camera's limit resolution according to the Rayleigh criterion are too loose, and the Dawes criterion is too strict. After considering the contrast test results on average, 5% is taken as the actual The feasible limit resolution benchmark contrast.

The main part of the resolution test is to obtain the target image contrast.

In order to obtain a good target imaging effect and facilitate theoretical analysis and calculation, the test target of this system breaks through the traditional target production mode. According to the needs, the light and dark stripes are made in the form of gradual gradation in a single direction, and the overall width and internal stripe width are fixed. , And the stripe width gradually narrows as the number of levels increases. During the test, after the CCD camera images the target, it takes a certain range of pixels in each light and dark fringe image as an image sample. The number of pixels selected in the calculation of each sample is set to 200×100, and then the sample contrast analysis is used to determine and then convert The process of calculating the resolution of the CCD camera is as follows:

(1) Determine the gray value Imax and the gray value Imin in the sample area through numerical judgment, and then calculate the average value I of the two as the boundary value of the light and dark stripes;

(2) Taking I as the boundary value, judge it as light or dark stripes line by line, and then calculate the average gray value Imax and Imin in the light and dark stripe area respectively; (3) Use the average gray value of pixels outside the target image area Calculate the background intensity Ib;

(4) Substitute Imax, Imin and Ib into formula (6) to calculate the overall contrast of this sample;

(5) Judge whether the sample is distinguishable by comparing the contrast of the sample with the reference contrast;

(6) After step-by-step judgment, until the limit resolution level reached by the image sample is confirmed, the system conversion calculates the overall resolution reached by the camera.

When manually interpreting the imaging resolution, it is required that the stripes in different directions in each unit on the target surface can be distinguished, so that this unit can be confirmed and distinguished. Therefore, in the automatic test, the target needs to be adjusted in two directions according to the current general standard and imaged separately, whichever is selected The calculated resolution value is used as the final test result.

3 Analysis of test results

For a certain type of CCD camera, according to two methods of manual interpretation test and automatic test, the corresponding resolution test data was obtained by using standard target and gradient grating target respectively. The resolution results obtained by the two test methods are shown in Table 1 and Table 2.



Through the comparative analysis of Table 1 and Table 2, it can be seen that the results of the two test methods are not much different, and the error does not exceed a set of resolution units, which belongs to the acceptable error range.

  4 Conclusion

The CCD camera resolution automatic test method proposed in this paper realizes the objective evaluation of the camera's technical indicators and avoids the influence of human factors. Moreover, this method can be adapted to various CCD cameras, corresponding to different types of cameras, as long as the image data acquisition interface and communication protocol are modified accordingly, the test can be completed conveniently and quickly, and the resolution result can be given, which is important for mass production of CCD The quality test of the camera has high practicality. (Author: Wang Feng, Gao Yafei)