High-precision line-scan digital camera splicing correction method

Abstract

The invention discloses a high-precision line-scan digital camera splicing correction method, particularly relates to the field of visual inspection, and comprises the steps of monitoring area division, data acquisition, data analysis and comprehensive analysis. According to the method, the proportional relation between the line frequency and the movement speed is corrected in real time through the dynamic parameter calibration system, image stretching and compression distortion is eliminated, accurate reduction of the physical size is guaranteed, the adaptive overlapping region expansion and feature matching optimization technology is adopted, pixel offset and gray jump caused by mechanical vibration are effectively restrained, the seamless splicing effect is achieved, and the image quality is improved. A laser calibration device and a high-precision guide rail are integrated to reduce the installation inclination error, a bilinear interpolation algorithm optimization and calibration maintenance mechanism is combined, the image detail reduction capability and the long-term stability of the system are comprehensively improved, and the imaging quality consistency of an industrial detection scene is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of visual detection technology, and more specifically, to a high-precision linear array camera stitching and correction method. Background Art

[0002] The existing technology mainly controls the uniform movement of the motion platform by presetting the line frequency parameters, and realizes image stitching by matching the feature points in the overlapping area of the field of view of adjacent cameras. The specific implementation includes three stages: first, the camera monitoring area is divided based on the theoretical line frequency speed ratio, then the conveyor belt speed is controlled by encoder feedback, and finally the SIFT feature matching algorithm is used to complete the image alignment and fusion.

[0003] However, this technology has systematic defects: first, the static matching model of line frequency and motion speed cannot adapt to the drift of the speed controller, resulting in irreversible stretching or compression distortion of the stitched image; second, the design that relies on a fixed overlapping area width makes it difficult to eliminate the feature point offset caused by mechanical vibration, and the fixed exposure parameters aggravate the grayscale mutation at the stitching seam; third, the inclination of the camera mounting plane and the straightness error of the guide rail will accumulate to form row and column misalignment artifacts; fourth, the calibration error caused by the contamination of the calibration plate is superimposed on the inefficient interpolation algorithm, resulting in the attenuation of the resolution of the stitched image and the loss of details. These problems seriously restrict the application of high-precision detection scenarios. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a high-precision line array camera stitching correction method, which solves the problems of insufficient dynamic matching accuracy of line frequency and motion speed, stitching misalignment caused by feature point offset, image deformation and resolution reduction caused by accumulation of mechanical installation errors and failure of calibration parameters proposed in the above-mentioned background technology through the following scheme.

[0005] To achieve the above object, the present invention provides the following technical solution: a high-precision linear array camera stitching correction method, comprising the following steps:

[0006] S1: Monitoring area division: used to determine the target linear array camera as the target area, divide the target area into sub-areas in a grid pattern, and the adjacent sub-areas include a 20%-30% overlap area, and mark each sub-area as 1, 2...n in sequence;

[0007] S2: Data collection: used for data collection for basic parameter verification of each sub-area, alignment check of overlapping areas, motion and installation error check, and resolution and calibration verification, and preprocessing of the collected data;

[0008] S3: Data analysis: By establishing a mathematical model, the data collected by S2 is analyzed to obtain the basic parameter matching, overlap alignment accuracy, motion and installation stability, and calibration and resolution consistency;

[0009] S4: Comprehensive analysis: Conduct comprehensive analysis based on the analysis results of S3, including the analysis of the matching degree of basic parameters, the analysis of the overlapping alignment accuracy, the analysis of the motion and installation stability, and the analysis of the calibration and resolution consistency.

[0010] Preferably, the data for basic parameter verification includes the line frequency of the line array camera, the actual speed of the motion platform, the actual physical length of the scanned object, and the total length of the stitched image; the data for overlapping area alignment inspection includes the overlapping pixel width of two adjacent frames of images, the X-direction offset of the feature points in adjacent images, the alignment error of the feature points after stitching, and the gray level difference at the stitching location; the data for motion and installation error inspection includes the installation tilt angle of the line array camera, the straightness error of the motion platform, the jitter of the line synchronization signal of the line array camera, and the global tilt angle of the stitched image; the data for resolution and calibration verification includes the calibrated pixel equivalent of the line array camera, the deviation between the actual physical size and the pixel calculation size, the line pair contrast of the resolution test chart, and the resolution degradation rate of the stitched image.

[0011] Preferably, in the basic parameter verification, the line frequency directly obtains the real-time setting value through the camera control software or the hardware configuration interface, the motion speed uses the real-time feedback data of the encoder or the motion control system, the actual length of the object is physically measured at both ends of the object to be measured using a high-precision scale, and the total length of the stitched image is statistically analyzed for the total number of horizontal pixels through an image processing tool.

[0012] Preferably, in the overlapping area alignment inspection, the overlapping pixel width is obtained by manually selecting the overlapping area in adjacent original images and counting the number of pixels, the feature point offset is calculated by the image comparison tool by selecting 3-5 feature points in the overlapping area and calculating the coordinate difference between two frames, the alignment error is measured by the residual offset of the feature points in the stitched image from the theoretical position, and the gray level difference uses the image analysis software to extract the gray level histogram of the overlapping area and calculate the mean difference or standard deviation.

[0013] Preferably, in the motion and installation error inspection, the installation tilt angle directly reads the X / Y axis angle values by closely attaching a digital inclinometer to the camera housing plane, the straightness error uses a laser interferometer to scan along the stroke of the motion platform and record the peak value of the lateral offset, the jitter of the synchronization signal is measured by connecting the camera line trigger signal line with an oscilloscope to measure the maximum time jitter value within 10 cycles, and the image tilt angle draws a reference line in the stitched image and uses an angle measurement tool to calculate the angle between it and the horizontal axis.

[0014] Preferably, the pixel equivalent in the resolution and calibration verification is calculated using a standard calibration plate through the ratio of a known physical spacing to a corresponding number of pixels, the size deviation is obtained by comparing the percentage error between the actual size of the calibration plate feature and the image measurement value, the line pair contrast is calculated by extracting a specified line pair area in a single frame image of the resolution test card using an MTF analysis tool, and the resolution drop rate is determined by comparing the contrast loss percentage of the same line pair in a single frame and in a stitched image.

[0015] Preferably, the basic parameter matching degree is specifically expressed as: E1 represents the basic parameter matching degree, F represents the line frequency of the linear array camera, V represents the actual speed of the motion platform, L real Indicates the actual physical length of the scanned object, Limg is the total length of the image after stitching.

[0016] Preferably, the overlap alignment accuracy is specifically expressed as: E2 represents the overlap alignment accuracy, W represents the overlap pixel width of two adjacent frames of images, Δx represents the X-direction offset of the feature point in the adjacent image, E represents the alignment error of the feature point after splicing, and D gray Indicates the grayscale difference at the splicing point.

[0017] Preferably, the motion and installation stability are specifically expressed as: E3 represents the motion and installation stability, θ represents the installation tilt angle of the linear array camera, δ represents the straightness error of the motion platform, J represents the line synchronization signal jitter of the linear array camera, and φ represents the global tilt angle of the stitched image.

[0018] Preferably, the calibration and resolution consistency are specifically expressed as: E4 indicates the consistency between calibration and resolution, P indicates the calibrated pixel equivalent of the line array camera, B indicates the deviation between the actual physical size and the calculated pixel size, C indicates the line pair contrast of the resolution test card, and R indicates the resolution drop rate of the stitched image.

[0019] Preferably, the basic parameter matching degree analysis is specifically expressed as follows: when E1>2%, it indicates that the matching degree between the line frequency F and the motion speed V is insufficient, resulting in the spliced image length Limg being inconsistent with the actual object length L real There is a significant deviation, which is caused by image stretching or compression due to speed controller calibration error or calibration parameter failure. It is necessary to calibrate the V / F ratio first and check the calibration parameter P in conjunction.

[0020] Preferably, the overlap alignment accuracy analysis is specifically expressed as follows: when E2>5%, it reflects that the feature point offset Δx or the residual error E in the overlap area is too large, accompanied by a grayscale difference D gray, indicating a scenario of motion platform jitter or sudden light change, it is necessary to increase the overlap width W to more than 10% of a single frame, optimize the feature matching algorithm, and enhance the exposure consistency control.

[0021] Preferably, the analysis of motion and installation stability is specifically expressed as: when E3 > 1.0, it is prompted that the mechanical installation tilt angle θ, the image tilt angle φ, the straightness error δ, or the synchronous jitter J jointly cause row-column misalignment. It is necessary to use laser to calibrate the camera installation plane, replace the high-precision linear guide rail, and optimize the anti-interference design of the trigger signal to reduce the composite error.

[0022] Preferably, the analysis of calibration and resolution consistency is specifically expressed as: when E4 > 10%, it indicates that the calibration plate size deviation B is out of tolerance or the resolution reduction rate R after splicing is too high. The root cause is calibration plate contamination, lens distortion, or interpolation algorithm defects. It is necessary to clean the calibration plate regularly, use bilinear interpolation instead of the nearest neighbor algorithm, and reduce the number of image resampling times to maintain the contrast C.

[0023] The technical effects and advantages of the present invention:

[0024] 1. By establishing a multi-dimensional data acquisition and mathematical model analysis mechanism, the present invention effectively improves the parameter matching accuracy of the linear array camera splicing system. By calibrating the dynamic proportional relationship between the line frequency and the motion speed in real time and jointly verifying the calibration parameters, the problem of image deformation caused by the speed controller error or parameter failure is eliminated, ensuring a high degree of consistency between the length of the spliced image and the actual physical size;

[0025] 2. The overlapping area optimization strategy of the present invention significantly improves the smoothness and alignment accuracy of image splicing. By increasing the width of the overlapping area, optimizing the feature matching algorithm, and controlling the exposure consistency, the offset of feature points and the gray jump caused by the motion platform jitter and sudden light change are effectively suppressed, achieving pixel-level alignment at the splicing point and reducing visual artifacts;

[0026] 3. Through the composite error compensation mechanism, the present invention comprehensively improves the system stability and imaging quality. By using laser to calibrate the camera installation plane, replacing the high-precision linear guide rail, and designing the anti-interference trigger signal, the influence of the composite error of mechanical tilt and line synchronization jitter is significantly reduced. At the same time, through the optimization of the bilinear interpolation algorithm and the calibration plate cleaning and maintenance process, the lens distortion and resolution loss are reduced, ensuring the calibration consistency and detail restoration ability of the spliced image. Brief Description of the Drawings

[0027] Figure 1 It is a schematic diagram of the overall structure of the present invention. Detailed Embodiment

[0028] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0029] This embodiment provides a high-precision linear array camera stitching and calibration method. Referring to Figure 1 as shown, it includes:

[0030] S1: Monitoring area division: Used to determine the target linear array camera as the target area, divide the target area into each sub-area in a grid pattern, the adjacent sub-areas include an overlapping area of 20%-30%, and sequentially label each sub-area as 1, 2... n.

[0031] S2: Data acquisition: Used to perform data acquisition on each sub-area for basic parameter verification, overlapping area alignment check, motion and installation error check, and resolution and calibration verification, and preprocess the acquired data.

[0032] The data for basic parameter verification includes the line frequency of the linear array camera, the actual speed of the motion platform, the actual physical length of the scanned object, and the total length of the stitched image; the data for overlapping area alignment check includes the overlapping pixel width of two adjacent frames of images, the X-direction offset of the feature points in adjacent images, the alignment error of the feature points after stitching, and the gray level difference at the stitching location; the data for motion and installation error check includes the installation tilt angle of the linear array camera, the straightness error of the motion platform, the jitter of the line synchronization signal of the linear array camera, and the global tilt angle of the stitched image; the data for resolution and calibration verification includes the calibrated pixel equivalent of the linear array camera, the deviation between the actual physical size and the pixel calculation size, the line pair contrast of the resolution test card, and the resolution reduction rate of the stitched image.

[0033] In the basic parameter verification, the line frequency directly obtains the real-time setting value through the camera control software or the hardware configuration interface, the motion speed uses the real-time feedback data of the encoder or the motion control system, the actual length of the object is physically measured at both ends of the object to be measured using a high-precision scale, and the total length of the stitched image is statistically analyzed for the total number of horizontal pixels through an image processing tool.

[0034] In the overlapping area alignment check, the overlapping pixel width is obtained by manually selecting the overlapping area in adjacent original images and counting the number of pixels. The feature point offset is calculated by the image comparison tool by selecting 3-5 feature points in the overlapping area and calculating the coordinate difference between two frames. The alignment error is measured by the residual offset between the feature points and the theoretical position in the stitched image. The gray level difference uses the image analysis software to extract the gray level histogram of the overlapping area and calculate the mean difference or standard deviation.

[0035] During the inspection of motion and installation errors, the installation tilt angle is directly read as the X / Y axis angle value by closely attaching a digital inclinometer to the plane of the camera housing. The straightness error is measured by scanning the travel of the motion platform with a laser interferometer and recording the peak value of the lateral offset. The jitter of the synchronization signal is measured by connecting an oscilloscope to the camera row trigger signal line to measure the maximum time jitter value within 10 cycles. For the image tilt angle, a reference line is drawn in the stitched image and an angle measurement tool is used to calculate the angle between it and the horizontal axis.

[0036] During the resolution and calibration verification, the pixel equivalent is calculated by the ratio of the known physical pitch to the corresponding number of pixels using a standard calibration board. The dimensional deviation is obtained by the percentage error of comparing the actual size of the calibration board features with the image measurement values. The line pair contrast is calculated by extracting the specified line pair area in a single-frame image of the resolution test chart using an MTF analysis tool. The resolution degradation rate is determined by comparing the percentage loss of contrast of the same line pair in single-frame and stitched images.

[0037] S3: Data analysis: Analyze the data collected in S2 by establishing a mathematical model to obtain the matching degree of basic parameters, the overlapping alignment accuracy, the motion and installation stability, and the calibration and resolution consistency.

[0038] The matching degree of basic parameters is specifically expressed as: E1 represents the matching degree of basic parameters, F represents the line frequency of the line array camera, V represents the actual speed of the motion platform, L real represents the actual physical length of the scanned object, and Limg represents the total length of the stitched image.

[0039] The overlapping alignment accuracy is specifically expressed as: E2 represents the overlapping alignment accuracy, W represents the overlapping pixel width of two adjacent frames of images, Δx represents the X-direction offset of the feature points in adjacent images, E represents the alignment error of the feature points after stitching, and D gray represents the gray level difference at the stitching location.

[0040] The motion and installation stability is specifically expressed as: E3 represents the motion and installation stability, θ represents the installation tilt angle of the line array camera, δ represents the straightness error of the motion platform, J represents the jitter of the line array camera's row synchronization signal, and φ represents the global tilt angle of the stitched image.

[0041] The calibration and resolution consistency is specifically expressed as: E4 represents the calibration and resolution consistency, P represents the calibrated pixel equivalent of the line array camera, B represents the deviation between the actual physical size and the pixel calculation size, C represents the line pair contrast of the resolution test chart, and R represents the resolution degradation rate of the stitched image.

[0042] P is used for physical size to pixel conversion, implicit in B.

[0043] S4: Comprehensive analysis: Comprehensive analysis is performed based on the analysis results of S3, including basic parameter matching analysis, overlap alignment accuracy analysis, motion and installation stability analysis, and calibration and resolution consistency analysis.

[0044] The basic parameter matching analysis is specifically expressed as follows: When E1>2%, it indicates that the matching degree between the line frequency F and the motion speed V is insufficient, resulting in the spliced image length Limg being inconsistent with the actual object length L real There is a significant deviation, which is caused by the speed controller calibration error or calibration parameter failure, resulting in image stretching or compression. It is necessary to calibrate the V / F ratio first and check the calibration parameter P in conjunction;

[0045] The specific expression of overlap alignment accuracy analysis is: when E2>5%, it reflects that the feature point offset Δx or residual error E in the overlap area is too large, accompanied by grayscale difference D gray , indicating that the motion platform is jittery or the illumination changes suddenly, it is necessary to increase the overlap width W to more than 10% of a single frame, optimize the feature matching algorithm and enhance the exposure consistency control;

[0046] The specific expression of motion and installation stability analysis is as follows: when E3>1.0, it indicates that the mechanical installation tilt angle θ, the image tilt angle φ and the straightness error δ or the synchronous jitter J together cause the row and column misalignment. It is necessary to use laser to calibrate the camera installation plane, replace the high-precision linear guide rail and optimize the trigger signal anti-interference design to reduce the compound error;

[0047] The calibration and resolution consistency analysis is specifically expressed as follows: when E4>10%, it indicates that the calibration plate size deviation B is out of tolerance or the resolution drop rate R after stitching is too high. The root cause is calibration plate contamination, lens distortion or interpolation algorithm defects. The calibration plate should be cleaned regularly, bilinear interpolation should be used instead of the nearest neighbor algorithm, and the number of image resampling times should be reduced to maintain the contrast C.

[0048] The present invention first divides the camera coverage area into sub-areas with 20%-30% overlap according to a grid and numbers them. Then, the line frequency is read by the camera software, the speed is obtained by the motion platform encoder, and the length of the object is measured by a high-precision ruler to complete the basic parameter verification. The image tool is used to analyze the pixel offset and grayscale difference in the overlapping area. The mechanical installation inclination angle and the motion platform straightness error are collected with the help of inclinometers, laser interferometers and other equipment. At the same time, a calibration plate is used to detect the pixel size deviation and the resolution drop rate. Then, the error indicators of four dimensions are calculated through a mathematical model, including speed matching, splicing alignment accuracy, motion stability and calibration consistency. Finally, corresponding correction measures are triggered according to the threshold, such as calibrating the speed controller, expanding the overlapping area, replacing the high-precision guide rail or upgrading the interpolation algorithm, so as to form a closed-loop control to achieve sub-pixel splicing accuracy.

[0049] Through the establishment of a multi-dimensional data acquisition and mathematical model analysis mechanism, the present invention effectively improves the parameter matching accuracy of the linear array camera stitching system. By calibrating the dynamic proportional relationship between the line frequency and the movement speed in real time and jointly verifying the calibration parameters, the problem of image deformation caused by the speed controller error or parameter failure is eliminated, ensuring a high degree of consistency between the length of the stitched image and the actual physical size.

[0050] The overlapping area optimization strategy of the present invention significantly improves the smoothness and alignment accuracy of image stitching. By increasing the width of the overlapping area, adopting feature matching algorithm optimization and exposure consistency control, the feature point offset and gray level jump caused by the movement platform jitter and sudden light change are effectively suppressed, achieving pixel-level alignment at the stitching position and reducing visual artifacts.

[0051] Through the composite error compensation mechanism, the present invention comprehensively improves the system stability and imaging quality. By using laser to calibrate the camera installation plane, replacing with a high-precision linear guide rail and designing an anti-interference trigger signal, the influence of the composite error of mechanical tilt and line synchronization jitter is significantly reduced. At the same time, through the optimization of the bilinear interpolation algorithm and the calibration plate cleaning and maintenance process, the lens distortion and resolution loss are reduced, ensuring the calibration consistency and detail restoration ability of the stitched image.

[0052] Secondly: In the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments are involved. For other structures, reference can be made to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other;

[0053] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A high-precision linear array camera stitching and calibration method, characterized in that, It includes the following steps: S1: Monitoring area division: Used to determine the target linear array camera as the target area, divide the target area into sub-areas in a grid pattern, with adjacent sub-areas including an overlapping area of 20%-30%, and sequentially label each sub-area as 1, 2... n; S2: Data acquisition: Used to perform data acquisition for basic parameter verification, overlapping area alignment check, motion and installation error check, and resolution and calibration verification of each sub-area, and preprocess the acquired data; S3: Data analysis: Analyze the data collected in S2 by establishing a mathematical model to obtain the basic parameter matching degree, overlapping alignment accuracy, motion and installation stability, and calibration and resolution consistency; S4: Comprehensive analysis: Conduct a comprehensive analysis based on the analysis results of S3, including basic parameter matching degree analysis, overlapping alignment accuracy analysis, motion and installation stability analysis, and calibration and resolution consistency analysis.

2. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The data for basic parameter verification includes the line frequency of the linear array camera, the actual speed of the motion platform, the actual physical length of the scanned object, and the total length of the stitched image; the data for overlapping area alignment check includes the overlapping pixel width of adjacent two frames of images, the X-direction offset of the feature points in adjacent images, the alignment error of the feature points after stitching, and the gray level difference at the stitching location; the data for motion and installation error check includes the installation tilt angle of the linear array camera, the straightness error of the motion platform, the jitter of the line synchronization signal of the linear array camera, and the global tilt angle of the image after stitching; the data for resolution and calibration verification includes the calibrated pixel equivalent of the linear array camera, the deviation between the actual physical size and the pixel calculation size, the line pair contrast of the resolution test card, and the resolution degradation rate of the image after stitching.

3. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The matching degree of the basic parameters is specifically expressed as: , E 1 represents the matching degree of the basic parameters, F represents the line frequency of the line array camera, V represents the actual speed of the moving platform, L real represents the actual physical length of the scanned object, L img The total length of the spliced image.

4. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The overlapping alignment accuracy is specifically expressed as: , E where 2 represents the overlapping alignment accuracy, W represents the overlapping pixel width of two adjacent frames of images, Δx represents the X-direction offset of feature points in adjacent images, E represents the alignment error of feature points after splicing, and D_gray represents the gray-level difference at the splicing position.

5. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The motion and installation stability are specifically expressed as: , E 3 represents the motion and installation stability, θ represents the installation tilt angle of the line array camera, δ represents the straightness error of the motion platform, J represents the jitter of the line synchronization signal of the line array camera, and ϕ represents the global tilt angle of the spliced image.

6. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The calibration and resolution consistency is specifically expressed as follows: , E 4 represents the calibration and resolution consistency, P represents the calibration pixel equivalent of the line array camera, B represents the deviation between the actual physical size and the pixel calculation size, C represents the line pair contrast of the resolution test card, and R represents the resolution degradation rate of the spliced image.

7. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The basic parameter matching degree analysis is specifically expressed as: E When 1>2%, it indicates that the matching degree between the line frequency F and the motion speed V is insufficient, resulting in a stitched image length L img There is a significant deviation from the actual object length L_real. The image is stretched or compressed due to speed controller calibration errors or calibration parameter failure. It is necessary to calibrate the V / F ratio first and check the calibration parameter P in conjunction.

8. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The specific expression of the overlapping alignment accuracy analysis is as follows: When E 2 > 5%, it reflects that the offset Δx or residual error E of the feature points in the overlapping area is too large, and at the same time, there is a gray-scale difference D gray , indicating a scenario of shaking of the moving platform or sudden change in illumination. It is necessary to increase the overlapping width W to more than 10% of a single frame, optimize the feature matching algorithm, and enhance the exposure consistency control.

9. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The analysis of the movement and installation stability is specifically expressed as: when E 3 > 1.0, it is prompted that the mechanical installation tilt angle θ, the image tilt angle ϕ, and the straightness error δ or the synchronous jitter J jointly cause row-column misalignment. It is necessary to use a laser to calibrate the camera installation plane, replace the high-precision linear guide rail, and optimize the anti-interference design of the trigger signal to reduce the composite error.

10. A high-precision linear array camera stitching and calibration method according to claim 1, characterized in that: The calibration and resolution consistency analysis is specifically expressed as: When E 4>10%, it indicates that the size deviation B of the calibration plate is out of tolerance or the resolution degradation rate R after splicing is too high. The root causes are calibration plate contamination, lens distortion, or interpolation algorithm defects. The calibration plate should be regularly cleaned, bilinear interpolation should be used to replace the nearest neighbor algorithm, and the number of image resamplings should be reduced to maintain the contrast C.

Images