Face recognition camera distortion test method and equipment

Abstract

The invention discloses a face recognition camera distortion test method and equipment. The distortion test method comprises the following steps: receiving a target pixel and a target adjustment factor of a to-be-tested camera; according to the target adjustment factor, the original distortion image and the distortion correction image, distortion parameters are determined, the original distortion image is an image which is not subjected to distortion correction, and the distortion correction image is an image which is subjected to distortion correction; and according to the target pixel and the distortion parameter, determining a test result of the to-be-tested camera, the test result being used for determining whether the to-be-tested camera is a qualified product. The distortion test method provided by the invention has the beneficial effects that the distortion parameter is automatically calculated, so that the quality of the face recognition camera product is verified, whether the face recognition camera product is qualified or not is automatically determined, the face recognition camera product leaving a factory is ensured to have high-quality production control, and the production efficiency is improved. And thus, the satisfaction degree of customers to products is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of computer vision image processing, and in particular to a method and device for testing distortion of a face recognition camera. Background Art

[0002] In the fields of high-end photography, scientific research, industrial inspection, etc., customers have extremely high requirements for the imaging quality of face recognition cameras, and distortion (geometric distortion) testing is one of the key elements to ensure that the product (face recognition camera) meets the requirements.

[0003] Therefore, in the product production process, distortion testing is one of the important links of quality control. The traditional face recognition camera distortion test is to draw some simple standard pictures manually, and then open the pictures before and after the distortion test with PhotoShop for manual measurement and manual calculation, and then compare and verify, and then screen out unqualified face recognition camera products. The disadvantages of manual measurement and manual calculation are high production costs, measurement and calculation vary from person to person, and there may be non-standard verification results, which affects the overall production quality and reduces customer satisfaction with the product; at the same time, when facing a large order volume, it requires a lot of labor costs and time costs, the efficiency is very low, and the test accuracy is not high, which does not meet the modern manufacturing industry's demand for efficient and high-quality products. Summary of the invention

[0004] In order to overcome the deficiencies of the prior art, the object of the present invention is to provide a face recognition camera distortion test method and device, which can solve the problems caused by manual measurement and manual calculation.

[0005] In order to solve the above problems, the present invention is implemented according to the following scheme:

[0006] A face recognition camera distortion test method is provided, including:

[0007] Receive target pixels and target adjustment factors of the camera to be tested;

[0008] Determining a distortion parameter according to the target adjustment factor, an original distortion map and a distortion correction map, wherein the original distortion map is a picture that has not been subjected to distortion correction, and the distortion correction map is a picture that has been subjected to distortion correction;

[0009] A test result of the camera to be tested is determined according to the target pixel and the distortion parameter, and the test result is used to determine whether the camera to be tested is a qualified product.

[0010] Compared with the prior art, the beneficial effects of the face recognition camera distortion test method provided by the present invention are as follows: by automatically calculating the distortion parameters, the quality of the face recognition camera product is verified to automatically determine whether the face recognition camera product is qualified, ensuring that the face recognition camera products leaving the factory have high-quality production control, thereby improving customer satisfaction with the product.

[0011] Optionally, the receiving of a target pixel and a target adjustment factor of the camera to be tested includes:

[0012] determining the received sensor pixel as the target pixel;

[0013] When the received adjustment factor is less than or equal to a preset threshold, determining the received adjustment factor as the target adjustment factor;

[0014] When the received adjustment factor is greater than the preset threshold, the adjustment factor is received again until the received adjustment factor is less than or equal to the preset threshold, and the received adjustment factor is determined as the target adjustment factor.

[0015] Optionally, the original distortion image is a picture randomly generated by a turtle library in Python language.

[0016] Optionally, the distortion correction image is a picture taken by the camera to be tested after distortion correction.

[0017] Optionally, before obtaining the distortion correction map, the method further includes:

[0018] Setting the parameters of the camera to be tested to optimal parameters;

[0019] The camera to be tested shoots with the optimal parameters to obtain the distortion correction image, and the distortion correction image has the same shape type as the original distortion image.

[0020] Optionally, determining the distortion parameter according to the target adjustment factor, the original distortion map and the distortion correction map includes:

[0021] Determining a first image parameter of the original distortion image;

[0022] determining a second image parameter of the distortion correction image;

[0023] The target adjustment factor, the first image parameter, and the second image parameter are calculated to obtain the distortion parameter.

[0024] Optionally, determining a test result of the camera to be tested according to the target pixel and the distortion parameter includes:

[0025] Determine a pixel type corresponding to the target pixel;

[0026] Determining a distortion parameter threshold according to the pixel type;

[0027] The test result is determined according to the distortion parameter and the distortion parameter threshold.

[0028] Optionally, determining the test result according to the distortion parameter and the distortion parameter threshold includes:

[0029] When the distortion parameter is less than the distortion parameter threshold, the test result is a qualified product;

[0030] When the distortion parameter is greater than or equal to the distortion parameter threshold, the test result is an unqualified product.

[0031] Optionally, also include:

[0032] Determine the number of qualified products and the number of unqualified products according to the test results of the camera to be tested;

[0033] The sum of the qualified product data and the number of unqualified products is determined as the number of test cameras;

[0034] A test report is generated based on the number of qualified products and the number of test cameras, and the test report is the product qualification rate.

[0035] A computer device is also provided, comprising a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the distortion testing method. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 The figure is a flow chart of the distortion testing method of the present invention. DETAILED DESCRIPTION

[0037] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0038] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and do not have to be used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0039] S1: Receive the target pixel and target adjustment factor of the camera to be tested, including:

[0040] Receive the sensor pixels and adjustment factors sent by the tester, determine the received sensor pixels as target pixels, and automatically check whether the received adjustment factors meet the preset threshold. When the received adjustment factors are less than or equal to the preset threshold, determine the received adjustment factors as the target adjustment factors. When the received adjustment factors are greater than the preset threshold, an error is reported, notifying the tester to resend the adjustment factors, and receiving the adjustment factors resent by the tester until the received adjustment factors are less than or equal to the preset threshold. Then determine the received adjustment factors as the target adjustment factors.

[0041] The number of sensor pixels is determined by the type of sensor used by the camera to be tested. The number of sensor pixels directly affects the image resolution and image quality of the camera. That is, the number of sensor pixels determines the image details and clarity that the camera to be tested can capture during the test.

[0042] The purpose of determining the target adjustment factor is to decouple the entire project and to better meet customer demand for products. In order to meet customers' basic requirements for products, the preset threshold of the present invention is 100,000. In the present invention, the smaller the value of the target adjustment factor, the higher the customer's requirements for the product. In the actual testing process, the size of the target adjustment factor can be adjusted according to the customer's requirements for the product. For example, if the customer's requirements for the product are twice as stringent, the target adjustment factor only needs to be 100,000÷2=50,000; if the customer's requirements for the product are 10 times more stringent, the target adjustment factor only needs to be 100,000÷10=10,000.

[0043] S2: Determine the distortion parameters according to the target adjustment factor, the original distortion map and the distortion correction map, wherein the original distortion map is a picture that has not been subjected to distortion correction, and the distortion correction map is a picture that has been subjected to distortion correction.

[0044] The original distortion image is a randomly generated image through the turtle library program fragment of the Python language. It can randomly generate standard images of triangle, rectangle, circle and other image shapes, and use the randomly generated standard image as the original distortion image that has not been corrected for distortion.

[0045] The distortion correction image is a picture taken by the camera to be tested after distortion correction. The camera to be tested is a face recognition camera product. Before using the camera to be tested to obtain the distortion correction image, in order to ensure the accuracy and reliability of the test results, it is necessary to set the parameters of the camera to be tested to the optimal parameters. Specifically, the ISP parameters of the camera to be tested are debugged to the optimal level, and the relevant parameters of the camera to be tested are set to normal mode.

[0046] Camera ISP (Image Signal Processor) refers to a chip or module that specializes in processing digital images. It is mainly used to convert analog signals output by digital camera image sensors into digital signals, and perform various enhancements and processing on them to improve image quality and enhance image clarity, color reproduction, dynamic range and depth perception. By debugging the ISP parameters of the camera to be tested to the optimal state, it can be ensured that the camera to be tested outputs high-quality distortion correction images during the test process, reducing distortion errors caused by image quality problems, thereby ensuring the accuracy and reliability of the test.

[0047] The relevant parameters of the camera to be tested are set to normal mode, that is, special functions such as night mode, HDR, face enhancement, etc. are not turned on during shooting. These special functions may affect the distortion of the obtained distortion correction image; by setting the relevant parameters of the camera to be tested to normal mode, it can be ensured that the test results of the camera to be tested are not affected by these special functions, thereby ensuring the accuracy and reliability of the test.

[0048] In order to ensure the effectiveness and reliability of the distortion test of the face recognition camera, it is necessary to ensure that the shape type of the original distortion map and the distortion correction map are the same, that is, after determining the shape type of the randomly generated image, the shape type of the distortion correction map taken by the camera to be tested is the same as the shape type of the randomly generated image.

[0049] If the shape types of the original distortion map and the distortion correction map are different, it is impossible to compare the original distortion map and the distortion correction map, and it is impossible to accurately judge the degree of distortion, that is, it is impossible to determine the distortion parameters; for example, if the standard image (original distortion map) is a triangle, and the image after the distortion test (distortion correction map) becomes a quadrilateral, then the test result will not be able to directly reflect the distortion of the triangle, and thus the effectiveness and reliability of the distortion test cannot be guaranteed.

[0050] According to the target adjustment factor, the original distortion map and the distortion correction map, the distortion parameters are determined, including:

[0051] Determine the image parameters of the original distortion image and the distortion correction image respectively, the image parameters are determined based on the image shape type, and the image parameters are specifically the side lengths of the image. For example, if the image shape type is a triangle, the image parameters are the lengths of the three sides, the first image parameters of the original distortion image are a, b, and c, and the second image parameters of the distortion correction image are aa, bb, and cc; if the image shape type is a quadrilateral, the image parameters are the lengths of the four sides, the first image parameters of the original distortion image are a, b, c, and d, and the second image parameters of the distortion correction image are aa, bb, cc, and dd; if the image shape type is a circle, the image parameters are the circumference of the circle, the first image parameter of the original distortion image is c, and the second image parameter of the distortion correction image is cc.

[0052] After determining the first image parameter of the original distortion image and the second image parameter of the distortion correction image, the target adjustment factor, the first image parameter and the second image parameter are calculated to obtain the distortion parameter. The specific calculation method is as follows:

[0053] If the shape type is a triangle, the calculation formula is as follows:

[0054] Distortion=(|(aa-a)|+|(bb-b)|+|(cc-c)|)×F×100%

[0055] Wherein, Distortion is the distortion parameter, a, b, c are the first-image parameters of the original distortion image, aa, bb, cc are the second-image parameters of the distortion correction image, and F is the target adjustment factor.

[0056] If the shape type is a quadrilateral, the calculation formula is as follows:

[0057] Distortion=(|(aa-a)|+|(bb-b)|+|(cc-c)|+|(dd-d)|)×F×100%

[0058] Wherein, Distortion is the distortion parameter, a, b, c, d are the first-image parameters of the original distortion image, aa, bb, cc, dd are the second-image parameters of the distortion correction image, and F is the target adjustment factor.

[0059] If the shape type is circular, the calculation formula is as follows:

[0060] Distortion=|(cc-c)|×F×100%

[0061] Wherein, Distortion is the distortion parameter, c is the first image parameter of the original distortion image, cc is the second image parameter of the distortion correction image, and F is the target adjustment factor.

[0062] S3: Determine the test result of the camera to be tested according to the target pixel and distortion parameter. The test result is used to determine whether the camera to be tested is a qualified product, including:

[0063] First, determine the pixel type corresponding to the target pixel, then determine the distortion parameter threshold according to the pixel type, and finally determine the test result according to the distortion parameter and the distortion parameter threshold. The specific process is as follows:

[0064] When the pixel type is CIF, the distortion parameter threshold is 1%. If the distortion parameter is less than 1%, the test result of the camera to be tested is a qualified product. If the distortion parameter is greater than or equal to 1%, the test result of the camera to be tested is an unqualified product.

[0065] When the pixel type is D1 or 720P, the distortion parameter threshold is 0.5%. If the distortion parameter is less than 0.5%, the test result of the camera to be tested is a qualified product. If the distortion parameter is greater than or equal to 0.5%, the test result of the camera to be tested is an unqualified product.

[0066] When the pixel type is 1080P, the distortion parameter threshold is 0.05%. If the distortion parameter is less than 0.05%, the test result of the camera to be tested is a qualified product. If the distortion parameter is greater than or equal to 0.05%, the test result of the camera to be tested is an unqualified product.

[0067] When the pixel type is above 3 million, the distortion parameter threshold is 0.001%. If the distortion parameter is less than 0.001%, the test result of the camera to be tested is a qualified product. If the distortion parameter is greater than or equal to 0.001%, the test result of the camera to be tested is an unqualified product.

[0068] When the pixel type is CIF, the corresponding resolution is 352*288 pixels; when the pixel type is D1, the corresponding resolution is 720*480 pixels; when the pixel type is 720P, the corresponding resolution is 1280*720 pixels; when the pixel type is 1080P, the corresponding resolution is 1920*1080 pixels; when the pixel type is more than 3 million, the corresponding resolution is 2048*1536 pixels or higher; it can be seen that different pixel types correspond to different resolutions.

[0069] High-resolution images contain more details and pixels and are more sensitive to distortion, which means that even small distortions may have a significant impact on image quality. Low-resolution images have fewer pixels and are relatively less sensitive to distortion. Therefore, when performing distortion tests on cameras with high resolutions corresponding to pixel types, the distortion test will be more stringent, that is, the distortion parameter threshold corresponding to the pixel type will be smaller, to ensure the accuracy of the test results.

[0070] After determining the test results of each camera to be tested in this test, the number of qualified products and the number of unqualified products are determined according to the test results of each camera to be tested. The number of qualified products and the number of unqualified products are added together to obtain the number of test cameras for this test, so as to automatically calculate the product qualification rate of this test, and the product qualification rate is used as a test report and fed back to the tester and the customer. The calculation formula of the product qualification rate is as follows:

[0071] R=(p÷s)×100%.

[0072] Among them, R is the product qualification rate, p is the number of qualified products, and s is the number of test cameras.

[0073] The present invention automatically generates test standard images (original distortion images) of various image shape types through the turtle library of Python language, and automatically measures the image parameters (first image parameters and second image parameters) of the image before and after distortion, and then automatically calculates the distortion parameters according to the target adjustment factor, and compares and verifies the distortion parameters with the distortion parameter threshold to determine whether each camera to be tested is a qualified product. Finally, the test report is automatically generated and fed back to the tester and the customer, ensuring that the face recognition camera products leaving the factory have high-quality production control, thereby improving customer satisfaction with the product.

[0074] The present invention also provides a computer device, comprising a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the above-mentioned distortion testing method.

[0075] The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0076] The memory can be used to store the computer program or module, and the processor implements various functions of the distortion test method by running or executing the computer program or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0077] The above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A face recognition camera distortion test method, characterized in that: include: Receive target pixels and target adjustment factors of the camera to be tested; Determining a distortion parameter according to the target adjustment factor, an original distortion map and a distortion correction map, wherein the original distortion map is a picture that has not been subjected to distortion correction, and the distortion correction map is a picture that has been subjected to distortion correction; A test result of the camera to be tested is determined according to the target pixel and the distortion parameter, and the test result is used to determine whether the camera to be tested is a qualified product.

2. A face recognition camera distortion testing method according to claim 1, characterized in that: The receiving of the target pixel and the target adjustment factor of the camera to be tested includes: determining the received sensor pixel as the target pixel; When the received adjustment factor is less than or equal to a preset threshold, determining the received adjustment factor as the target adjustment factor; When the received adjustment factor is greater than the preset threshold, the adjustment factor is received again until the received adjustment factor is less than or equal to the preset threshold, and the received adjustment factor is determined as the target adjustment factor.

3. The method for testing distortion of a face recognition camera according to claim 1, characterized in that: The original distortion image is a randomly generated image using the turtle library in Python.

4. The method for testing distortion of a face recognition camera according to claim 1, characterized in that: The distortion correction image is a picture taken by the camera to be tested after distortion correction.

5. A face recognition camera distortion testing method according to claim 4, characterized in that: Before obtaining the distortion correction map, the method further includes: Setting the parameters of the camera to be tested to optimal parameters; The camera to be tested shoots with the optimal parameters to obtain the distortion correction image, and the distortion correction image has the same shape type as the original distortion image.

6. The method for testing distortion of a face recognition camera according to claim 1, characterized in that: Determining the distortion parameter according to the target adjustment factor, the original distortion map and the distortion correction map includes: Determining a first image parameter of the original distortion image; determining a second image parameter of the distortion correction image; The target adjustment factor, the first image parameter, and the second image parameter are calculated to obtain the distortion parameter.

7. A face recognition camera distortion testing method according to claim 6, characterized in that: Determining the test result of the camera to be tested according to the target pixel and the distortion parameter includes: Determine a pixel type corresponding to the target pixel; Determining a distortion parameter threshold according to the pixel type; The test result is determined according to the distortion parameter and the distortion parameter threshold.

8. The method for testing distortion of a face recognition camera according to claim 7, characterized in that: The determining the test result according to the distortion parameter and the distortion parameter threshold comprises: When the distortion parameter is less than the distortion parameter threshold, the test result is a qualified product; When the distortion parameter is greater than or equal to the distortion parameter threshold, the test result is an unqualified product.

9. A face recognition camera distortion testing method according to claim 8, characterized in that: Also includes: Determine the number of qualified products and the number of unqualified products according to the test results of the camera to be tested; The sum of the qualified product data and the number of unqualified products is determined as the number of test cameras; A test report is generated based on the number of qualified products and the number of test cameras, and the test report is the product qualification rate.

10. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the distortion testing method as described in any one of claims 1 to 9.

Images