Stereoscopic vision three-dimensional measurement system

Abstract

The invention discloses a stereoscopic vision three-dimensional measurement system, relates to the technical field of three-dimensional measurement, and aims to solve the technical problems that a current traditional detection method often depends on manual visual inspection or uses a simple measurement tool, the methods are time-consuming and labor-consuming, and the accuracy and consistency of a detection result are difficult to guarantee. Comprising the following steps: S1, image acquisition; a high-resolution camera is used, the resolution of the camera is not lower than 4096 * 3072, the camera has automatic focusing and optical anti-shaking functions, an advanced image sensor is arranged, the average color error does not exceed 3 delta E, the dynamic range is not smaller than 12 gears, and images of the surface of the countersunk bolt in the diameter direction are clearly captured. The camera can automatically adjust the focal length and shooting parameters according to different bolt sizes and detection requirements so as to ensure that the clearest image is obtained, and in the shooting process, the focal length adjustment of the camera can be carried out according to the following formula. The method has the advantage of improving the measurement accuracy.

Description

Technical Field

[0001] The present invention relates to the field of three-dimensional measurement technology, and more specifically, to a stereoscopic vision three-dimensional measurement system. Background Art

[0002] Stereoscopic 3D measurement is an advanced measurement technology that imitates the way human eyes observe the world and uses the principles of computer vision and precision measurement to obtain 3D information of objects. The working principle of stereoscopic 3D measurement is mainly based on the principle of parallax and triangulation. It uses one or more observation systems (usually cameras) to take 2D images of objects from different angles or at different times. Then, by using the corresponding points or feature points in these images, calculating the parallax between them and combining the triangular geometric relationship, the 3D information of the object, such as height, distance and shape, can be obtained.

[0003] At present, in the process of assembling precision parts, the installation quality of countersunk bolts is directly related to the stability and reliability of the entire assembly. Traditional detection methods often rely on manual visual inspection or the use of simple measuring tools. These methods are not only time-consuming and labor-intensive, but also difficult to ensure the accuracy and consistency of the detection results. The detection results rely on the subjective experience of the inspectors. Since the subjective judgments of different inspectors often differ, it is difficult to ensure the installation quality of countersunk bolts in precision parts. In view of this, we propose a stereoscopic vision 3D measurement system. Summary of the invention

[0004] The purpose of the present invention is to provide a stereoscopic vision three-dimensional measurement system to solve the technical problem that current traditional detection methods often rely on manual visual inspection or the use of simple measuring tools. These methods are not only time-consuming and labor-intensive, but also difficult to ensure the accuracy and consistency of the detection results.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a stereoscopic vision three-dimensional measurement system, comprising the following steps:

[0006] S1, image acquisition;

[0007] Use a high-resolution camera with a resolution of no less than 4096×3072, autofocus and optical image stabilization, and an advanced image sensor with an average color error of no more than 3ΔE and a dynamic range of no less than 12 stops to clearly capture the image of the surface of the countersunk bolt along the diameter direction;

[0008] During the shooting process, the camera can automatically adjust the focal length and shooting parameters according to different bolt sizes and detection requirements to ensure the clearest image. During the shooting process, the camera can automatically adjust the focal length and shooting parameters according to different bolt sizes and detection requirements to ensure the clearest image.

[0009] The camera's focal length can be adjusted according to the following formula:

[0010]

[0011] Where f is the focal length, D is the bolt diameter, S is the sensor size, and L is the shooting distance;

[0012] The lighting device provides uniform and stable lighting. The light intensity can be adjusted within 100-10000 lux, and the color temperature can be adjusted within 2700K-6500K. It is composed of multiple independently adjustable light sources to meet different shooting needs.

[0013] The intelligent control system can automatically adjust the lighting parameters within 0.1 seconds according to the camera position and shooting angle, and has an automatic brightness adjustment function with a response time of no more than 0.05, ensuring that the image quality is not affected by changes in lighting conditions;

[0014] The lighting device can also automatically switch between strong light mode and weak light mode according to the ambient light conditions to adapt to different shooting environments;

[0015] The light intensity can be adjusted according to the following formula:

[0016]

[0017] Where I is the light intensity, k is the light constant, d is the distance from the light source to the bolt, L is the distance from the light source to the camera, θ is the light angle, and α is the ambient light influence coefficient;

[0018] The camera is accurately aimed at the countersunk bolt through the positioning device for shooting, with the positioning accuracy not exceeding 0.01mm, the angle adjustment accuracy not exceeding 0.1°, and the position adjustment accuracy within 0.05mm;

[0019] Adopt high-precision mechanical structure and advanced control system to achieve fast and accurate positioning operation;

[0020] Equipped with precision displacement sensors and angle sensors, the measurement accuracy is 0.005mm, the angle measurement accuracy is less than 0.05°, data is collected at a frequency of 100Hz, and the camera position and angle changes are monitored in real time;

[0021] The positioning device can also exchange data with an external positioning system to improve the accuracy and stability of positioning;

[0022] The coordinate transformation of positioning can be performed using the following matrix formula:

[0023]

[0024] Where (x, y, z) is the coordinate of the positioning device, R is the rotation matrix, T is the translation vector, n is the number of external interference factors, ω i is the weight coefficient of the i-th interference factor, (x i ,y i , z i ) is the coordinate of the i-th interference factor;

[0025] S2, image processing;

[0026] S2.1, image preprocessing;

[0027] Using a multi-scale filtering algorithm, the signal-to-noise ratio of the denoising effect is improved by no less than 10dB.

[0028] The features of countersunk bolts are highlighted by contrast enhancement and edge sharpening. The image contrast is increased by no less than 50% after contrast enhancement, and the edge clarity is increased by more than 30% after edge sharpening.

[0029] It has image correction function with a correction accuracy within 0.1 pixel and color correction. The color difference after color correction does not exceed 2ΔE.

[0030] It has image segmentation function, and the segmentation accuracy is not less than 95%.

[0031] During image preprocessing, images can also be compressed and encoded to reduce the amount of data for subsequent processing and increase processing speed;

[0032] Image denoising can be performed using the following formula:

[0033]

[0034] Among them I in is the input image, I out is the output image, ω i，j，k is the weight coefficient, which is adjusted according to different scales k and noise types, and m is the number of scales;

[0035] S2.2, feature extraction;

[0036] The Canny edge detection method and texture analysis technology are used to extract the diameter, contour, and texture features of the countersunk bolts. The edge detection accuracy is within 0.2 pixels, and the texture analysis accuracy is not less than 90%;

[0037] The diameter extraction uses edge detection and fitting algorithms to determine the diameter by fitting the edge points with the least square method. The error of the diameter extraction does not exceed 0.05mm.

[0038] The contour extraction adopts the eight-neighborhood contour tracking algorithm, and the accuracy of contour extraction is not less than 90%;

[0039] Texture extraction uses gray-level co-occurrence matrix algorithm, and the accuracy of texture extraction is not less than 85%;

[0040] Using a multi-scale feature extraction algorithm, the feature richness is increased by no less than 30%;

[0041] The center axis is calculated using the least squares fitting method, and the error does not exceed 0.05mm;

[0042] The formula for feature extraction can be expressed as:

[0043]

[0044] Where F is the feature vector, gl, k is the feature extraction function, which is defined according to different feature types l and scales k. in is the input image, n is the number of feature types, m is the number of scales, and λl is the weight coefficient of the lth feature type;

[0045] A plurality of measuring points are formed on the top surface of the countersunk bolt in the diameter direction. First, the positioning device aims the high-resolution camera at the center of the countersunk bolt to take an initial image.

[0046] During the alignment process, laser positioning technology and high-precision sensors are used to ensure that the camera coincides with the center axis of the countersunk bolt, with an error of no more than 0.005mm;

[0047] At the same time, the positioning device will also calibrate the initial position of the camera to ensure that the coordinate system of the camera is consistent with the actual coordinate system of the countersunk bolt;

[0048] The positioning device adopts high-precision three-dimensional positioning technology, which can achieve fast and accurate positioning operations;

[0049] During the positioning process, the positioning device will also monitor the position changes of the camera in real time to ensure that the positioning accuracy is always within the specified range;

[0050] Then, taking the center of the countersunk bolt as a reference, the camera is rotated by the positioning device at preset angle intervals to photograph the surface of the countersunk bolt at different angles to establish multiple measurement points;

[0051] The angle interval can be adjusted according to the size of the countersunk bolt and the detection accuracy requirements, and the minimum angle interval does not exceed 5°;

[0052] During the shooting process, the positioning device will automatically adjust the light intensity and angle of the lighting device according to the position and shooting angle of the camera to ensure a clear image at each measuring point;

[0053] The positioning device adopts advanced motion control technology, which can achieve fast and accurate rotation operation;

[0054] During the rotation process, the positioning device will also monitor the angle changes of the camera in real time to ensure that the angle accuracy always remains within the specified range;

[0055] Measure multiple measurement points. Let the total number of measurement points be N. For the i-th measurement point, let the coordinates of a point on the surface be (x i ,y i , z i ), the center coordinates of the countersunk hole end face are (x0, y0, z0), then the distance from the measuring point to the countersunk hole end face is:

[0056]

[0057] Calculate the center axis vector of the countersunk bolt at the measuring point The perpendicular vector The angle is θ i , can be obtained through the vector dot product formula Calculate, where ∈ is a very small positive number to avoid the denominator being zero;

[0058] Comprehensively analyze the measurement results of multiple measurement points to determine whether the countersunk bolts are tightened;

[0059] First calculate the average distance Average Angle

[0060] The tolerance range can be adjusted according to different bolt types, sizes and installation requirements. For example, for a bolt with a diameter of 10 mm, the distance tolerance range is [D-ΔD, D+ΔD], and the angle tolerance range is [θ0-Δθ, θ0+Δθ], where D is the bolt diameter, ΔD is the diameter tolerance, θ0 is the vertical angle, and Δθ is the angle tolerance;

[0061] if and It is judged that the countersunk bolt is tightened;

[0062] If it is out of tolerance, it is judged that the countersunk bolt is not tightened;

[0063] At the same time, considering factors such as bolt material, size, and installation environment, the judgment criteria and algorithms are automatically adjusted within 0.5 seconds;

[0064] It has fault diagnosis function with an accuracy rate of not less than 90%;

[0065] The formula for analysis and judgment can be expressed as:

[0066]

[0067] Where S is the bolt tightening state, n is the number of influencing factors, ωi is the weight coefficient of the i-th influencing factor, f i is the function of the i-th influencing factor, and γ is the judgment threshold;

[0068] S3, result output;

[0069] The detection results are displayed in real time on the display screen with a resolution of no less than 1920×1080. The displayed content includes the countersunk bolt image, detection results and related parameters. It can be zoomed in, zoomed out and rotated. The response time of the operation does not exceed 0.2 seconds.

[0070] Adopt high brightness and high contrast display technology, with brightness ranging from 100-500cd / m 2 Internal adjustment, contrast ratio can be adjusted within 1000:1-5000:1;

[0071] It has an alarm function, the response time is no more than 0.1 seconds, and it has a touch screen function, and the response accuracy is within 0.5mm;

[0072] The display screen can also display real-time inspection progress and statistical information, including the number of inspected bolts, the number of qualified bolts, the number of unqualified bolts, etc., to facilitate users to understand the inspection status;

[0073] Use a printer to print the test report, with a print resolution of no less than 1200dpi, capable of printing QR codes or barcodes, with a generation time of no more than 0.5 seconds;

[0074] With network printing function, the response time does not exceed 1 second;

[0075] Printed reports include detailed bolt information, test results, and image screenshots;

[0076] Use data storage devices to store test data for subsequent query and analysis, with a storage capacity of no less than 1TB;

[0077] Using a database management system, the data backup interval does not exceed 1 hour, and the recovery time does not exceed 10 minutes;

[0078] It has data encryption function, the strength of encryption algorithm is above AES-256, and it can transmit data wirelessly with external devices, the transmission speed is not less than 50Mbps, and the transmission stability is above 95%;

[0079] Adopting distributed storage technology, data redundancy is not less than 2, and access response time is not more than 0.5 seconds;

[0080] Possess data mining capabilities with an accuracy rate of no less than 80%;

[0081] The formula for data storage can be expressed as:

[0082]

[0083] Where D in is the input data, D stored is the stored data, E is the encryption and storage function, n is the number of data optimization factors, and w i is the weight coefficient of the i-th optimization factor, h i is the function of the i-th optimization factor, is the average distance, is the average angle.

[0084] The present invention collects the status of the countersunk bolt after installation through image acquisition, and pre-processes and extracts features of the acquired image, and combines the algorithm formula to calculate and determine the distance and angle between a single point on the surface of the countersunk bolt and the end face of the countersunk hole, so as to comprehensively analyze and determine whether the countersunk bolt is tightened. There is no need for manual visual inspection or the use of simple measuring tools. Calculation and analysis are performed through objective data, which is conducive to avoiding measurement errors caused by subjective experience and judgment of detection personnel, thereby improving measurement accuracy.

[0085] Preferably, the autofocus system of the high-resolution camera is optimized using a deep learning algorithm to quickly and accurately focus according to different bolt surface features and shooting environments, with a focus accuracy within 0.01 mm and an autofocus response time of no more than 0.05 seconds;

[0086] At the same time, the camera has intelligent scene recognition function, which automatically adjusts the shooting parameters according to the different bolt types and installation scenarios detected to obtain the best image quality.

[0087] Preferably, the color temperature adjustment of the lighting device is monitored in real time using a high-precision color temperature sensor, and the adjustment accuracy is within 30K;

[0088] The lighting device automatically selects the best color temperature combination according to the reflection characteristics of different bolt materials to highlight the characteristics of the bolts and improve the clarity and contrast of the image.

[0089] Preferably, the data interaction between the positioning device and the external positioning system adopts high-speed wireless communication technology, with a data transmission rate of not less than 100Mbps and an interaction frequency of not less than 60Hz;

[0090] The positioning device can receive high-precision coordinate information from the external positioning system in real time and accurately adjust the position of the camera to ensure that the camera is always aligned with the center position of the countersunk bolt, and the positioning error does not exceed 0.008mm.

[0091] Preferably, the image compression algorithm in the image preprocessing adopts the image compression technology based on deep learning, and the compression ratio is not less than 15:1 under the premise of ensuring the image quality;

[0092] The compression algorithm automatically learns the features of different bolt images and efficiently compresses the images, reducing the amount of data and processing time for subsequent processing.

[0093] Preferably, the multi-scale feature extraction algorithm in the feature extraction adopts a multi-scale feature extraction method based on a convolutional neural network to automatically learn the features of bolts at different scales, and the feature extraction accuracy is not less than 92%;

[0094] The algorithm can effectively extract the edge, texture, shape and other features of the bolt, providing more accurate and rich information for subsequent analysis and judgment.

[0095] Preferably, during the photographing of multiple measuring points on the top surface of the countersunk bolt, the illumination intensity of the illumination device is adjusted using an intelligent illumination control algorithm, and the response time does not exceed 0.04 seconds;

[0096] The algorithm can automatically adjust the lighting intensity and angle according to the camera's shooting angle and the position of the measurement point, ensuring that the image of each measurement point has good lighting uniformity and contrast.

[0097] Preferably, when comprehensively analyzing and judging whether the countersunk bolt is tightened, the bolt material factors considered include physical parameters such as hardness, elastic modulus, yield strength, and characteristics such as thermal expansion coefficient and corrosion resistance of the material;

[0098] By establishing a relationship model between material parameters and tightening status, the tightening status of bolts made of different materials can be judged more accurately, thereby improving the accuracy and reliability of detection.

[0099] Preferably, the display screen used in the result output uses a high-resolution, high-color-saturation OLED display screen, with a resolution of not less than 2560×1440 and a color saturation of not less than 90%;

[0100] The display screen has wide viewing angle and fast response characteristics, and can clearly display the test results and bolt images, making it convenient for users to observe and analyze.

[0101] Preferably, the data encryption algorithm of the data storage device adopts multi-layer encryption technology, including a combination of symmetric encryption and asymmetric encryption;

[0102] The data storage device also has data backup and recovery functions. The backup time interval does not exceed 30 minutes, and the recovery time does not exceed 8 minutes, ensuring the security and reliability of the test data.

[0103] Compared with the prior art, the present invention has the following beneficial effects:

[0104] 1. The present invention collects the state of the countersunk bolt after installation through image acquisition, and pre-processes and extracts features of the collected images, and combines the algorithm formula to calculate and determine the distance and angle between a single point on the surface of the countersunk bolt and the end face of the countersunk hole, so as to comprehensively analyze and determine whether the countersunk bolt is tightened. There is no need for manual visual inspection or the use of simple measuring tools. Calculation and analysis are performed through objective data, which is conducive to avoiding measurement errors caused by subjective experience and judgment of detection personnel, thereby improving measurement accuracy, and solving the problem that traditional detection methods often rely on manual visual inspection or the use of simple measuring tools. These methods are not only time-consuming and labor-intensive, but also difficult to ensure the accuracy and consistency of detection results.

[0105] 2. The present invention also forms multiple measuring points in the diameter direction of the top surface of the countersunk bolt, and comprehensively analyzes the measurement results of the multiple measuring points to more comprehensively reflect the actual state of the bolt. This multi-measurement point method can reduce misjudgment caused by local abnormalities or measurement errors and improve the accuracy of judgment.

[0106] 3. The present invention also measures the distance and angle of multiple measurement points and calculates the average distance and average angle, making the judgment basis more stable and reliable. Compared with a single measurement value, the averaging process can reduce the impact of individual abnormal values ​​on the results and more accurately reflect the overall tightening state of the bolt. DETAILED DESCRIPTION

[0107] In order to facilitate those skilled in the art to understand the technical solution of the present invention, the technical solution of the present invention is now further described.

[0108] A stereoscopic vision three-dimensional measurement system comprises the following steps:

[0109] S1, image acquisition;

[0110] Use a high-resolution camera with a resolution of no less than 4096×3072, autofocus and optical image stabilization, and an advanced image sensor with an average color error of no more than 3ΔE and a dynamic range of no less than 12 stops to clearly capture the image of the surface of the countersunk bolt along the diameter direction;

[0111] During the shooting process, the camera can automatically adjust the focal length and shooting parameters according to different bolt sizes and detection requirements to ensure the clearest image. During the shooting process, the camera can automatically adjust the focal length and shooting parameters according to different bolt sizes and detection requirements to ensure the clearest image.

[0112] The camera's focal length can be adjusted according to the following formula:

[0113]

[0114] Where f is the focal length, D is the bolt diameter, S is the sensor size, and L is the shooting distance;

[0115] The lighting device provides uniform and stable lighting. The light intensity can be adjusted within 100-10000 lux, and the color temperature can be adjusted within 2700K-6500K. It is composed of multiple independently adjustable light sources to meet different shooting needs.

[0116] The intelligent control system can automatically adjust the lighting parameters within 0.1 seconds according to the camera position and shooting angle, and has an automatic brightness adjustment function with a response time of no more than 0.05, ensuring that the image quality is not affected by changes in lighting conditions;

[0117] The lighting device can also automatically switch between strong light mode and weak light mode according to the ambient light conditions to adapt to different shooting environments;

[0118] The light intensity can be adjusted according to the following formula:

[0119]

[0120] Where I is the light intensity, k is the light constant, d is the distance from the light source to the bolt, L is the distance from the light source to the camera, θ is the light angle, and α is the ambient light influence coefficient;

[0121] The camera is accurately aimed at the countersunk bolt through the positioning device for shooting, with the positioning accuracy not exceeding 0.01mm, the angle adjustment accuracy not exceeding 0.1°, and the position adjustment accuracy within 0.05mm;

[0122] Adopt high-precision mechanical structure and advanced control system to achieve fast and accurate positioning operation;

[0123] Equipped with precision displacement sensors and angle sensors, the measurement accuracy is 0.005mm, the angle measurement accuracy is less than 0.05°, data is collected at a frequency of 100Hz, and the camera position and angle changes are monitored in real time;

[0124] The positioning device can also exchange data with an external positioning system to improve the accuracy and stability of positioning;

[0125] The coordinate transformation of positioning can be performed using the following matrix formula:

[0126]

[0127] Where (x, y, z) is the coordinate of the positioning device, R is the rotation matrix, T is the translation vector, n is the number of external interference factors, ω i is the weight coefficient of the i-th interference factor, (x i ,y i , zi ) is the coordinate of the i-th interference factor;

[0128] S2, image processing;

[0129] S2.1, image preprocessing;

[0130] Using a multi-scale filtering algorithm, the signal-to-noise ratio of the denoising effect is improved by no less than 10dB.

[0131] The features of countersunk bolts are highlighted by contrast enhancement and edge sharpening. The image contrast is increased by no less than 50% after contrast enhancement, and the edge clarity is increased by more than 30% after edge sharpening.

[0132] It has image correction function with a correction accuracy within 0.1 pixel and color correction. The color difference after color correction does not exceed 2ΔE.

[0133] It has image segmentation function, and the segmentation accuracy is not less than 95%.

[0134] During image preprocessing, images can also be compressed and encoded to reduce the amount of data for subsequent processing and increase processing speed;

[0135] Image denoising can be performed using the following formula:

[0136]

[0137] Among them I in is the input image, I out is the output image, ω i，j，k is the weight coefficient, which is adjusted according to different scales k and noise types, and m is the number of scales;

[0138] S2.2, feature extraction;

[0139] The Canny edge detection method and texture analysis technology are used to extract the diameter, contour, and texture features of the countersunk bolts. The edge detection accuracy is within 0.2 pixels, and the texture analysis accuracy is not less than 90%;

[0140] The diameter extraction uses edge detection and fitting algorithms to determine the diameter by fitting the edge points with the least square method. The error of the diameter extraction does not exceed 0.05mm.

[0141] The contour extraction adopts the eight-neighborhood contour tracking algorithm, and the accuracy of contour extraction is not less than 90%;

[0142] Texture extraction uses gray-level co-occurrence matrix algorithm, and the accuracy of texture extraction is not less than 85%;

[0143] Using a multi-scale feature extraction algorithm, the feature richness is increased by no less than 30%;

[0144] The center axis is calculated using the least squares fitting method, and the error does not exceed 0.05mm;

[0145] The formula for feature extraction can be expressed as:

[0146]

[0147] Where F is the feature vector, gl, k is the feature extraction function, which is defined according to different feature types l and scales k. in is the input image, n is the number of feature types, m is the number of scales, and λl is the weight coefficient of the lth feature type;

[0148] A plurality of measuring points are formed on the top surface of the countersunk bolt in the diameter direction. First, the positioning device aims the high-resolution camera at the center of the countersunk bolt to take an initial image.

[0149] During the alignment process, laser positioning technology and high-precision sensors are used to ensure that the camera coincides with the center axis of the countersunk bolt, with an error of no more than 0.005mm;

[0150] At the same time, the positioning device will also calibrate the initial position of the camera to ensure that the coordinate system of the camera is consistent with the actual coordinate system of the countersunk bolt;

[0151] The positioning device adopts high-precision three-dimensional positioning technology, which can achieve fast and accurate positioning operations;

[0152] During the positioning process, the positioning device will also monitor the position changes of the camera in real time to ensure that the positioning accuracy is always within the specified range;

[0153] Then, taking the center of the countersunk bolt as a reference, the camera is rotated by the positioning device at preset angle intervals to photograph the surface of the countersunk bolt at different angles to establish multiple measurement points;

[0154] The angle interval can be adjusted according to the size of the countersunk bolt and the detection accuracy requirements, and the minimum angle interval does not exceed 5°;

[0155] During the shooting process, the positioning device will automatically adjust the light intensity and angle of the lighting device according to the position and shooting angle of the camera to ensure a clear image at each measuring point;

[0156] The positioning device adopts advanced motion control technology, which can achieve fast and accurate rotation operation;

[0157] During the rotation process, the positioning device will also monitor the angle changes of the camera in real time to ensure that the angle accuracy always remains within the specified range;

[0158] Measure multiple measurement points. Let the total number of measurement points be N. For the i-th measurement point, let the coordinates of a point on the surface be (x i ,y i , z i ), the center coordinates of the countersunk hole end face are (x0, y0, z0), then the distance from the measuring point to the countersunk hole end face is:

[0159]

[0160] Calculate the center axis vector of the countersunk bolt at the measuring point The perpendicular vector The angle is θ i , can be obtained through the vector dot product formula Calculate, where ∈ is a very small positive number to avoid the denominator being zero;

[0161] Comprehensively analyze the measurement results of multiple measurement points to determine whether the countersunk bolts are tightened;

[0162] First calculate the average distance Average Angle

[0163] The tolerance range can be adjusted according to different bolt types, sizes and installation requirements. For example, for a bolt with a diameter of 10 mm, the distance tolerance range is [D-ΔD, D+ΔD], and the angle tolerance range is [θ0-Δθ, θ0+Δθ], where D is the bolt diameter, ΔD is the diameter tolerance, θ0 is the vertical angle, and Δθ is the angle tolerance;

[0164] if and It is judged that the countersunk bolt is tightened;

[0165] If it is out of tolerance, it is judged that the countersunk bolt is not tightened;

[0166] At the same time, considering factors such as bolt material, size, and installation environment, the judgment criteria and algorithms are automatically adjusted within 0.5 seconds;

[0167] It has fault diagnosis function with an accuracy rate of not less than 90%;

[0168] The formula for analysis and judgment can be expressed as:

[0169]

[0170] Where S is the bolt tightening state, n is the number of influencing factors, ω i is the weight coefficient of the i-th influencing factor, f i is the function of the i-th influencing factor, and γ is the judgment threshold;

[0171] S3, result output;

[0172] The detection results are displayed in real time on the display screen with a resolution of no less than 1920×1080. The displayed content includes the countersunk bolt image, detection results and related parameters. It can be zoomed in, zoomed out and rotated. The response time of the operation does not exceed 0.2 seconds.

[0173] Adopt high brightness and high contrast display technology, with brightness ranging from 100-500cd / m 2 Internal adjustment, contrast ratio can be adjusted within 1000:1-5000:1;

[0174] It has an alarm function, the response time is no more than 0.1 seconds, and it has a touch screen function, and the response accuracy is within 0.5mm;

[0175] The display screen can also display real-time inspection progress and statistical information, including the number of inspected bolts, the number of qualified bolts, the number of unqualified bolts, etc., to facilitate users to understand the inspection status;

[0176] Use a printer to print the test report, with a print resolution of no less than 1200dpi, capable of printing QR codes or barcodes, with a generation time of no more than 0.5 seconds;

[0177] With network printing function, the response time does not exceed 1 second;

[0178] Printed reports include detailed bolt information, test results, and image screenshots;

[0179] Use data storage devices to store test data for subsequent query and analysis, with a storage capacity of no less than 1TB;

[0180] Using a database management system, the data backup interval does not exceed 1 hour, and the recovery time does not exceed 10 minutes;

[0181] It has data encryption function, the strength of encryption algorithm is above AES-256, and it can transmit data wirelessly with external devices, the transmission speed is not less than 50Mbps, and the transmission stability is above 95%;

[0182] Adopting distributed storage technology, data redundancy is not less than 2, and access response time is not more than 0.5 seconds;

[0183] Possess data mining capabilities with an accuracy rate of no less than 80%;

[0184] The formula for data storage can be expressed as:

[0185]

[0186] Where D in is the input data, D storedis the stored data, E is the encryption and storage function, n is the number of data optimization factors, and w i is the weight coefficient of the i-th optimization factor, h i is the function of the i-th optimization factor, is the average distance, is the average angle.

[0187] The high-resolution camera's autofocus system is optimized using a deep learning algorithm to quickly and accurately focus according to different bolt surface features and shooting environments, with a focus accuracy of less than 0.01mm and an autofocus response time of no more than 0.05 seconds.

[0188] At the same time, the camera has intelligent scene recognition function, which automatically adjusts the shooting parameters according to the different bolt types and installation scenes detected to obtain the best image quality. The color temperature adjustment of the lighting device adopts high-precision color temperature sensor for real-time monitoring, and the adjustment accuracy is within 30K;

[0189] The lighting device automatically selects the best color temperature combination according to the reflection characteristics of different bolt materials to highlight the characteristics of the bolts and improve the clarity and contrast of the image. The data interaction between the positioning device and the external positioning system adopts high-speed wireless communication technology, with a data transmission rate of not less than 100Mbps and an interaction frequency of not less than 60Hz.

[0190] The positioning device can receive high-precision coordinate information from the external positioning system in real time and accurately adjust the position of the camera to ensure that the camera is always aligned with the center of the countersunk bolt, with a positioning error of no more than 0.008 mm. The image compression algorithm in image preprocessing uses image compression technology based on deep learning. Under the premise of ensuring image quality, the compression ratio is not less than 15:1.

[0191] The compression algorithm automatically learns the features of different bolt images and efficiently compresses the images, reducing the amount of data and processing time for subsequent processing.

[0192] In the embodiment of the present invention, the multi-scale feature extraction algorithm in the feature extraction adopts a multi-scale feature extraction method based on a convolutional neural network to automatically learn the features of bolts at different scales, and the feature extraction accuracy is not less than 92%;

[0193] The algorithm can effectively extract the edge, texture, shape and other features of the bolt, providing more accurate and rich information for subsequent analysis and judgment. During the shooting of multiple measurement points on the top surface of the countersunk bolt, the lighting intensity of the lighting device is adjusted using an intelligent lighting control algorithm, with a response time of no more than 0.04 seconds.

[0194] The algorithm can automatically adjust the light intensity and angle according to the camera's shooting angle and the position of the measurement point, ensuring that the image of each measurement point has good light uniformity and contrast. When comprehensively analyzing whether the countersunk bolts are tightened, the bolt material factors considered include physical parameters such as hardness, elastic modulus, yield strength, and material properties such as thermal expansion coefficient and corrosion resistance.

[0195] By establishing a relationship model between material parameters and tightening status, the tightening status of bolts of different materials can be judged more accurately, improving the accuracy and reliability of detection. The display screen in the result output uses a high-resolution, high-color saturation OLED display screen with a resolution of no less than 2560×1440 and a color saturation of no less than 90%;

[0196] The display screen has wide viewing angle and fast response characteristics, and can clearly display the test results and bolt images, which is convenient for users to observe and analyze. The data encryption algorithm of the data storage device adopts multi-layer encryption technology, including a combination of symmetric encryption and asymmetric encryption;

[0197] The data storage device also has data backup and recovery functions. The backup time interval does not exceed 30 minutes, and the recovery time does not exceed 8 minutes, ensuring the security and reliability of the test data.

[0198] The embodiments of the present invention disclose preferred embodiments, but are not limited thereto. A person skilled in the art can easily understand the spirit of the present invention based on the above embodiments and make different extensions and changes. However, as long as they do not deviate from the spirit of the present invention, they are all within the protection scope of the present invention.

Claims

1. A stereoscopic three-dimensional measurement system, characterized in that: The following steps are involved: S1, image acquisition; Use a high-resolution camera with a resolution of no less than 4096×3072, autofocus and optical image stabilization, and an advanced image sensor with an average color error of no more than 3ΔE and a dynamic range of no less than 12 stops to clearly capture the image of the surface of the countersunk bolt along the diameter direction; During the shooting process, the camera can automatically adjust the focal length and shooting parameters according to different bolt sizes and detection requirements to ensure the clearest image. During the shooting process, the camera can automatically adjust the focal length and shooting parameters according to different bolt sizes and detection requirements to ensure the clearest image. The camera's focal length can be adjusted according to the following formula: Where f is the focal length, D is the bolt diameter, S is the sensor size, and L is the shooting distance; The lighting device provides uniform and stable lighting. The light intensity can be adjusted within 100-10000 lux, and the color temperature can be adjusted within 2700K-6500K. It is composed of multiple independently adjustable light sources to meet different shooting needs. The intelligent control system can automatically adjust the lighting parameters within 0.1 seconds according to the camera position and shooting angle, and has an automatic brightness adjustment function with a response time of no more than 0.05, ensuring that the image quality is not affected by changes in lighting conditions; The lighting device can also automatically switch between strong light mode and weak light mode according to the ambient light conditions to adapt to different shooting environments; The light intensity can be adjusted according to the following formula: Where I is the light intensity, k is the light constant, d is the distance from the light source to the bolt, L is the distance from the light source to the camera, θ is the light angle, and α is the ambient light influence coefficient; The camera is accurately aimed at the countersunk bolt through the positioning device for shooting, with the positioning accuracy not exceeding 0.01mm, the angle adjustment accuracy not exceeding 0.1°, and the position adjustment accuracy within 0.05mm; Adopt high-precision mechanical structure and advanced control system to achieve fast and accurate positioning operation; Equipped with precision displacement sensors and angle sensors, the measurement accuracy is 0.005mm, the angle measurement accuracy is less than 0.05°, data is collected at a frequency of 100Hz, and the camera position and angle changes are monitored in real time; The positioning device can also exchange data with an external positioning system to improve the accuracy and stability of positioning; The coordinate transformation of positioning can be performed using the following matrix formula: Where (x, y, z) is the coordinate of the positioning device, R is the rotation matrix, T is the translation vector, n is the number of external interference factors, ω i is the weight coefficient of the i-th interference factor, (x i ,y i , z i ) is the coordinate of the i-th interference factor; S2, image processing; S2.1, image preprocessing; Using a multi-scale filtering algorithm, the signal-to-noise ratio of the denoising effect is improved by no less than 10dB. The features of countersunk bolts are highlighted by contrast enhancement and edge sharpening. The image contrast is increased by no less than 50% after contrast enhancement, and the edge clarity is increased by more than 30% after edge sharpening. It has image correction function with a correction accuracy within 0.1 pixel and color correction. The color difference after color correction does not exceed 2ΔE. It has image segmentation function, and the segmentation accuracy is not less than 95%. During image preprocessing, images can also be compressed and encoded to reduce the amount of data for subsequent processing and increase processing speed; Image denoising can be performed using the following formula: Among them I in is the input image, I out is the output image, ω i，j，k is the weight coefficient, which is adjusted according to different scales k and noise types, and m is the number of scales; S2.2, feature extraction; The Canny edge detection method and texture analysis technology are used to extract the diameter, contour, and texture features of the countersunk bolts. The edge detection accuracy is within 0.2 pixels, and the texture analysis accuracy is not less than 90%; The diameter extraction uses edge detection and fitting algorithms to determine the diameter by fitting the edge points with the least square method. The error of the diameter extraction does not exceed 0.05mm. The contour extraction adopts the eight-neighborhood contour tracking algorithm, and the accuracy of contour extraction is not less than 90%; Texture extraction uses gray-level co-occurrence matrix algorithm, and the accuracy of texture extraction is not less than 85%; Using a multi-scale feature extraction algorithm, the feature richness is increased by no less than 30%; The center axis is calculated using the least squares fitting method, and the error does not exceed 0.05mm; The formula for feature extraction can be expressed as: Where F is the feature vector, gl, k is the feature extraction function, which is defined according to different feature types l and scales k. in is the input image, n is the number of feature types, m is the number of scales, and λl is the weight coefficient of the lth feature type; A plurality of measuring points are formed on the top surface of the countersunk bolt in the diameter direction. First, the positioning device aims the high-resolution camera at the center of the countersunk bolt to take an initial image. During the alignment process, laser positioning technology and high-precision sensors are used to ensure that the camera coincides with the center axis of the countersunk bolt, with an error of no more than 0.005mm; At the same time, the positioning device will also calibrate the initial position of the camera to ensure that the coordinate system of the camera is consistent with the actual coordinate system of the countersunk bolt; The positioning device adopts high-precision three-dimensional positioning technology, which can achieve fast and accurate positioning operations; During the positioning process, the positioning device will also monitor the position changes of the camera in real time to ensure that the positioning accuracy is always within the specified range; Then, taking the center of the countersunk bolt as a reference, the camera is rotated by the positioning device at preset angle intervals to photograph the surface of the countersunk bolt at different angles to establish multiple measurement points; The angle interval can be adjusted according to the size of the countersunk bolt and the detection accuracy requirements, and the minimum angle interval does not exceed 5°; During the shooting process, the positioning device will automatically adjust the light intensity and angle of the lighting device according to the position and shooting angle of the camera to ensure a clear image at each measuring point; The positioning device adopts advanced motion control technology, which can achieve fast and accurate rotation operation; During the rotation process, the positioning device will also monitor the angle changes of the camera in real time to ensure that the angle accuracy always remains within the specified range; Measure multiple measurement points. Let the total number of measurement points be N. For the i-th measurement point, let the coordinates of a point on the surface be (x i ,y i , z i ), the center coordinates of the countersunk hole end face are (x0, y0, z0), then the distance from the measuring point to the countersunk hole end face is: Calculate the center axis vector of the countersunk bolt at the measuring point The perpendicular vector The angle is θ i , can be obtained through the vector dot product formula Calculate, where ∈ is a very small positive number to avoid the denominator being zero; Comprehensively analyze the measurement results of multiple measurement points to determine whether the countersunk bolts are tightened; First calculate the average distance Average Angle The tolerance range can be adjusted according to different bolt types, sizes and installation requirements. For example, for a bolt with a diameter of 10 mm, the distance tolerance range is [D-ΔD, D+ΔD], and the angle tolerance range is [θ0-Δθ, θ0+Δθ], where D is the bolt diameter, ΔD is the diameter tolerance, θ0 is the vertical angle, and Δθ is the angle tolerance; if and It is judged that the countersunk bolt is tightened; If it is out of tolerance, it is judged that the countersunk bolt is not tightened; At the same time, considering factors such as bolt material, size, and installation environment, the judgment criteria and algorithms are automatically adjusted within 0.5 seconds; It has fault diagnosis function with an accuracy rate of not less than 90%; The formula for analysis and judgment can be expressed as: Where S is the bolt tightening state, n is the number of influencing factors, ω i is the weight coefficient of the i-th influencing factor, f i is the function of the i-th influencing factor, and γ is the judgment threshold; S3, result output; The detection results are displayed in real time on the display screen with a resolution of no less than 1920×1080. The displayed content includes the countersunk bolt image, detection results and related parameters. It can be zoomed in, zoomed out and rotated. The response time of the operation does not exceed 0.2 seconds. Adopt high brightness and high contrast display technology, with brightness ranging from 100-500cd / m 2 Internal adjustment, contrast ratio can be adjusted within 1000:1-5000:1; It has an alarm function, the response time is no more than 0.1 seconds, and it has a touch screen function, and the response accuracy is within 0.5mm; The display screen can also display real-time inspection progress and statistical information, including the number of inspected bolts, the number of qualified bolts, the number of unqualified bolts, etc., to facilitate users to understand the inspection status; Use a printer to print the test report, with a print resolution of no less than 1200dpi, capable of printing QR codes or barcodes, with a generation time of no more than 0.5 seconds; With network printing function, the response time does not exceed 1 second; Printed reports include detailed bolt information, test results, and image screenshots; Use data storage devices to store test data for subsequent query and analysis, with a storage capacity of no less than 1TB; Using a database management system, the data backup interval does not exceed 1 hour, and the recovery time does not exceed 10 minutes; It has data encryption function, the strength of encryption algorithm is above AES-256, and it can transmit data wirelessly with external devices, the transmission speed is not less than 50Mbps, and the transmission stability is above 95%; Adopting distributed storage technology, data redundancy is not less than 2, and access response time is not more than 0.5 seconds; Possess data mining capabilities with an accuracy rate of no less than 80%; The formula for data storage can be expressed as: Where D in is the input data, D stored is the stored data, E is the encryption and storage function, n is the number of data optimization factors, and w i is the weight coefficient of the i-th optimization factor, h i is the function of the i-th optimization factor, is the average distance, is the average angle.

2. A stereoscopic vision three-dimensional measurement system according to claim 1, characterized in that: The high-resolution camera's autofocus system is optimized using a deep learning algorithm to quickly and accurately focus according to different bolt surface features and shooting environments, with a focus accuracy of less than 0.01mm and an autofocus response time of no more than 0.05 seconds. At the same time, the camera has intelligent scene recognition function, which automatically adjusts the shooting parameters according to the different bolt types and installation scenarios detected to obtain the best image quality.

3. A stereoscopic vision three-dimensional measurement system according to claim 2, characterized in that: The color temperature adjustment of the lighting device uses a high-precision color temperature sensor for real-time monitoring, and the adjustment accuracy is within 30K; The lighting device automatically selects the best color temperature combination according to the reflection characteristics of different bolt materials to highlight the characteristics of the bolts and improve the clarity and contrast of the image.

4. A stereoscopic vision three-dimensional measurement system according to claim 3, characterized in that: The data exchange between the positioning device and the external positioning system adopts high-speed wireless communication technology, with a data transmission rate of not less than 100Mbps and an interaction frequency of not less than 60Hz; The positioning device can receive high-precision coordinate information from the external positioning system in real time and accurately adjust the position of the camera to ensure that the camera is always aligned with the center position of the countersunk bolt, and the positioning error does not exceed 0.008mm.

5. A stereoscopic vision three-dimensional measurement system according to claim 4, characterized in that: The image compression algorithm in image preprocessing adopts image compression technology based on deep learning. Under the premise of ensuring image quality, the compression ratio is not less than 15:1; The compression algorithm automatically learns the features of different bolt images and efficiently compresses the images, reducing the amount of data and processing time for subsequent processing.

6. The stereoscopic vision three-dimensional measurement system according to claim 5, characterized in that: The multi-scale feature extraction algorithm in feature extraction adopts a multi-scale feature extraction method based on convolutional neural network to automatically learn the features of bolts at different scales, and the feature extraction accuracy is not less than 92%; The algorithm can effectively extract the edge, texture, shape and other features of the bolt, providing more accurate and rich information for subsequent analysis and judgment.

7. The stereoscopic vision three-dimensional measurement system according to claim 6, characterized in that: During the shooting of multiple measurement points on the top surface of the countersunk bolt, the lighting intensity of the lighting device is adjusted using an intelligent lighting control algorithm with a response time of no more than 0.04 seconds; The algorithm can automatically adjust the illumination intensity and angle according to the camera’s shooting angle and the position of the measurement point, ensuring that the image of each measurement point has good illumination uniformity and contrast.

8. The stereoscopic vision three-dimensional measurement system according to claim 7, characterized in that: When comprehensively analyzing and judging whether the countersunk bolts are tightened, the bolt material factors to be considered include physical parameters such as hardness, elastic modulus, yield strength, and characteristics such as thermal expansion coefficient and corrosion resistance of the material; By establishing a relationship model between material parameters and tightening status, the tightening status of bolts made of different materials can be judged more accurately, thereby improving the accuracy and reliability of detection.

9. The stereoscopic vision three-dimensional measurement system according to claim 1, characterized in that: The display screen used in the result output uses a high-resolution, high-color-saturation OLED display screen, with a resolution of no less than 2560×1440 and a color saturation of no less than 90%; The display screen has wide viewing angle and fast response characteristics, and can clearly display the test results and bolt images, making it convenient for users to observe and analyze.

10. The stereoscopic vision three-dimensional measurement system according to claim 1, characterized in that: The data encryption algorithm of the data storage device adopts multi-layer encryption technology, including a combination of symmetric encryption and asymmetric encryption; The data storage device also has data backup and recovery functions. The backup time interval does not exceed 30 minutes, and the recovery time does not exceed 8 minutes, ensuring the security and reliability of the test data.

Images