Method for realizing cross-image size measurement through coordinate transformation

Abstract

The invention relates to the technical field of size measurement, in particular to a method for achieving cross-image size measurement through coordinate transformation, and the method comprises the steps: controlling the movement time and position of a lens through a servo motor, so as to guarantee the positioning of the lens during each shooting; a product is moved to a proper position through the feeding mechanism, and it is ensured that a measured object and a lens are kept horizontal and parallel, so that measurement errors caused by angle deviation are reduced; a high-definition and high-speed camera is used for shooting images on the sliding rail according to a preset path in a regional mode, and overlapping areas of a certain proportion are reserved between adjacent shooting areas; a coordinate conversion algorithm is adopted; according to the method, a high-definition and high-speed camera, a high-precision servo motor and an advanced coordinate conversion algorithm are combined, accurate measurement of a large number of sizes can be completed within extremely short time, and the real-time detection requirement on an industrial automatic production line is met. Through an adaptive correction algorithm, the problem of angular deviation of images obtained at different visual angles is solved.

Description

Technical Field

[0001] The present application relates to the technical field of dimension measurement, and in particular to a method for realizing cross-image dimension measurement through coordinate transformation. Background Art

[0002] In modern industrial production and quality control, dimensional measurement is an important part of ensuring product quality and consistency. Traditional dimensional measurement methods usually rely on contact measurement tools or single image analysis, which are inefficient, inaccurate, and difficult to adapt to objects with complex geometries. With the continuous improvement of industrial automation, there is an increasing demand for non-contact, high-precision, high-speed dimensional measurement that can adapt to a variety of different shapes and sizes of products.

[0003] Among the existing non-contact measurement technologies, although some methods based on image analysis can provide certain measurement capabilities, they are often limited by factors such as shooting angles and changes in lighting conditions, resulting in unstable and inaccurate measurement results. In addition, when dealing with dimensional measurements across multiple images, traditional methods are difficult to achieve high-precision unified measurements due to the lack of an effective coordinate conversion mechanism. At the same time, when faced with high-speed production lines that require real-time detection, the speed and response time of existing technologies are also insufficient. Summary of the invention

[0004] In view of the above, it is necessary to provide a method for cross-image size measurement through coordinate transformation.

[0005] The present application provides a method for realizing cross-image size measurement through coordinate transformation, comprising: Use servo motors to control the timing and position of lens movement to ensure that the lens is positioned correctly for each shot; The product is moved to a suitable position through the feeding mechanism to ensure that the measuring object is horizontal and parallel to the lens, thereby reducing the measurement error caused by angle deviation; Use high-definition high-speed cameras to capture images in different areas along a predetermined path on the slide rail, and retain a certain proportion of overlapping areas between adjacent shooting areas to ensure image continuity and integrity, and support subsequent image stitching; By adopting the coordinate transformation algorithm, the measurement points in multiple images are unified into a common coordinate system, realizing high-precision and high-speed cross-image dimension measurement, which can meet the needs of real-time detection on industrial production lines.

[0006] In one possible implementation, fixed features in the overlapping area of ​​adjacent images are further used for calibration, and mathematical operations (such as affine transformation or perspective transformation) are used to ensure accurate alignment of coordinate systems between different images, which significantly improves the measurement accuracy and can adapt to measurement objects of different shapes and sizes.

[0007] In one possible implementation, the center coordinates of the overlapping area features of each image are calculated and used as reference points to achieve unified alignment of all measurement point coordinates. This method not only improves the reliability of measurement, but also simplifies the multi-image stitching process, which is a function not available in existing technologies.

[0008] In one possible implementation, an adaptive correction algorithm is used during the coordinate conversion process so that images acquired at different viewing angles can be accurately converted into the same coordinate system, thus solving the angle deviation problem that is difficult to handle with traditional methods. This method is particularly suitable for measuring objects with complex geometric shapes.

[0009] In one possible implementation, the method completes the measurement of more than 100 dimensions within 1 second, meeting the demand for high-speed measurement in the manufacturing process while providing measurement speed and accuracy far exceeding existing technologies. It is suitable for real-time detection on industrial automation production lines, greatly improving production efficiency.

[0010] In one possible implementation, the measurement area is decomposed into multiple image areas, and the image magnification is increased several times for high-precision measurement. This method is particularly suitable for the strict requirements of the precision manufacturing industry. This method can capture tiny size differences and provides the possibility for measurement accuracy that cannot be easily achieved with existing technologies.

[0011] In one possible implementation, this method can greatly improve the flexibility and adaptability of the measurement system without increasing additional hardware costs, allowing users to adjust the measurement range and accuracy according to specific needs, thereby optimizing resource utilization and return on investment.

[0012] In a possible implementation, an autofocus function is also included, which can automatically adjust the lens focus according to the distance of the measured object to ensure that images taken at different positions have optimal clarity, thereby improving the reliability of the measurement results.

[0013] In one possible implementation, the method is configured with an ambient light compensation mechanism to maintain stable imaging quality under different lighting conditions and avoid measurement errors caused by light changes. It is particularly suitable for application scenarios that require measurements in changing environments.

[0014] In one possible implementation, it also includes a built-in automatic defect detection function, which analyzes image data based on deep learning algorithms, quickly identifies and marks product parts that do not meet quality standards, assists manual review, and improves product quality control.

[0015] In the above-mentioned method of realizing cross-image dimension measurement through coordinate transformation, this method combines high-definition high-speed cameras, high-precision servo motors and advanced coordinate transformation algorithms, and can complete accurate measurement of a large number of dimensions in a very short time, meeting the real-time detection needs of industrial automation production lines. The problem of angle deviation of images obtained from different viewing angles is solved through the adaptive correction algorithm, which is particularly suitable for measuring objects with complex geometric shapes.

[0016] Supports dynamic adjustment of image magnification. Users can select different magnification modes according to specific needs, flexibly responding to various types of measurement tasks without adding additional hardware costs. Built-in ambient light compensation mechanism and autofocus function ensure that products at different lighting conditions and distances can obtain clear and stable images, avoiding measurement errors caused by light changes or distance differences.

[0017] The use of multi-threaded parallel processing technology and hardware acceleration technology (such as FPGA, GPU) greatly shortens the time of image processing and data analysis, and greatly improves production efficiency.

[0018] The automatic defect detection module based on deep learning can quickly identify and mark product parts that do not meet quality standards, assist manual review, and improve product quality control. It also supports detection of multiple defect types and customizes detection rules and thresholds according to user needs.

[0019] The intuitive operation experience provided by the software interface allows users to easily adjust measurement parameters and save configuration templates for easy use next time, thereby optimizing resource utilization and return on investment. The reliability of the system is enhanced by introducing backup motors or sensors and closed-loop control algorithms; the application of visual assistance systems and air-floating guide rails / magnetic levitation technology further improves the positioning accuracy and stability of the feeding mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0021] Figure 1 It is a schematic diagram of grabbing measurement areas by regions during camera movement in a method for realizing cross-image size measurement by coordinate transformation in one embodiment of the present application; Figure 2 In one embodiment of the present application, multiple images are stitched into one schematic diagram; Figure 3 It is a schematic diagram of traditional image stitching in one embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0023] See also Figure 1-Figure 3 This embodiment provides a technical solution: a method for realizing cross-image size measurement through coordinate transformation, comprising: Use servo motors to control the timing and position of lens movement to ensure that the lens is positioned correctly for each shot; The product is moved to a suitable position through the feeding mechanism to ensure that the measuring object is horizontal and parallel to the lens, thereby reducing the measurement error caused by angle deviation; Use high-definition high-speed cameras to capture images in different areas along a predetermined path on the slide rail, and retain a certain proportion of overlapping areas between adjacent shooting areas to ensure image continuity and integrity, and support subsequent image stitching; By adopting the coordinate transformation algorithm, the measurement points in multiple images are unified into a common coordinate system, realizing high-precision and high-speed cross-image dimension measurement, which can meet the needs of real-time detection on industrial production lines.

[0024] In some embodiments, fixed features in the overlapping area of ​​adjacent images are further utilized for calibration, and mathematical operations (such as affine transformation or perspective transformation) are used to ensure accurate alignment of coordinate systems between different images, thereby significantly improving measurement accuracy and being able to adapt to measurement objects of different shapes and sizes.

[0025] By calculating the center coordinates of the overlapping area characteristics of each image and using it as a reference point, the coordinates of all measurement points are aligned uniformly. This method not only improves the reliability of measurement, but also simplifies the multi-image stitching process, which is a function not available in existing technologies. In the coordinate conversion process, an adaptive correction algorithm is used to accurately convert images acquired at different viewing angles into the same coordinate system, solving the angle deviation problem that is difficult to handle with traditional methods, and is particularly suitable for measuring objects with complex geometric shapes.

[0026] Furthermore, it can complete the measurement of more than 100 dimensions within 1 second, meeting the needs of high-speed measurement in the manufacturing process, while providing measurement speed and accuracy far exceeding existing technologies. It is suitable for real-time detection on industrial automation production lines, greatly improving production efficiency.

[0027] The measurement area is decomposed into multiple image areas, and the image magnification is increased several times for high-precision measurement. This method is particularly suitable for the strict requirements of the precision manufacturing industry. This method can capture tiny size differences and provides the possibility for measurement accuracy that cannot be easily achieved with existing technologies.

[0028] This method can greatly improve the flexibility and adaptability of the measurement system without increasing additional hardware costs, allowing users to adjust the measurement range and accuracy according to specific needs, thereby optimizing resource utilization and return on investment.

[0029] It also includes an autofocus function that automatically adjusts the lens focus according to the distance of the measurement object, ensuring that images taken at different positions have optimal clarity, thereby improving the reliability of measurement results.

[0030] This method is equipped with an ambient light compensation mechanism to maintain stable imaging quality under different lighting conditions and avoid measurement errors caused by light changes. It is particularly suitable for application scenarios that require measurements in changing environments.

[0031] It also includes a built-in automatic defect detection function, which analyzes image data based on deep learning algorithms, quickly identifies and marks product parts that do not meet quality standards, assists manual review, and improves product quality control.

[0032] It should be noted that the servo motor control and lens positioning: Use a high-precision servo motor (such as the Panasonic MG series) to control the position of the lens. The motor has an angular displacement accuracy of ±0.05 degrees, ensuring that the lens can be accurately positioned to the predetermined position for each shot. The motor response time does not exceed 20 milliseconds. Through a fast feedback control system, positioning adjustments can be completed in a very short time to ensure the timeliness and accuracy of shooting. The lens movement range is ±30 degrees horizontally and ±20 degrees vertically, which is suitable for shooting needs at a variety of different angles. The motor is equipped with a built-in encoder to monitor the position of the lens in real time to ensure the stability of positioning.

[0033] Working principle: The servo motor drives the lens to move along a predetermined path by receiving instructions from the control system. The encoder continuously feeds back the actual position of the lens, and the control system makes fine adjustments based on the feedback information to ensure that the lens is always in the specified position. Before each shot, the system automatically calibrates the lens position to ensure that it is completely consistent with the preset coordinates, thereby reducing measurement deviations caused by mechanical errors.

[0034] Furthermore, by adding spare motors or sensors, the reliability and fault tolerance of the system can be improved. A closed-loop control algorithm (such as PID control) can be used to further improve positioning accuracy and response speed.

[0035] In some embodiments, the feeding mechanism uses a high-precision linear guide system (such as the THK LR series) with a positioning accuracy of up to ±0.1 mm. The guide system can accurately position the product 300 mm in front of the lens during each measurement cycle. The feeding platform is equipped with multiple fine-tuning screws, allowing the operator to fine-tune the product in the X, Y, and Z directions to ensure that the product remains parallel and level with the lens. The system is also equipped with a laser rangefinder to monitor the distance between the product and the lens in real time to ensure that both are always within the optimal shooting distance range.

[0036] Working principle: The feeding mechanism is driven by a motor to feed the product from the feed port to the measuring area in sequence. After each feeding, the system automatically detects the position of the product and corrects it through the fine-tuning screw to ensure that the product remains parallel and level with the lens. The laser rangefinder monitors the distance between the product and the lens in real time. Once a deviation is found, the system will immediately adjust the height of the feeding platform to ensure a clear focus when shooting.

[0037] Furthermore, a visual assistance system is introduced to use a camera to capture the product outline, automatically identify and adjust the product position, and further improve positioning accuracy. Air-floating guide rails or magnetic levitation technology are used to reduce friction and increase feeding speed and stability.

[0038] High-definition high-speed camera and image acquisition: Use industrial-grade CMOS cameras with a resolution of up to 2048x2048 pixels (such as the Basler ace series) and a frame rate of up to 1000 frames per second to ensure that clear images can be captured even at high speeds.

[0039] The camera is mounted on a slide rail and captures images in different areas along a predetermined path. The overlap between adjacent shooting areas is 20%, that is, each image covers 20% of the area of ​​the previous image, ensuring image continuity and integrity. The slide rail system is driven by a stepper motor and can move the camera with a step accuracy of 0.1 mm, ensuring that the camera is positioned accurately for each shot.

[0040] Working principle: The camera moves along the slide rail and takes images of different areas one by one according to the preset path. Each time it takes a picture, the camera stops briefly to ensure the stability of the image. A 20% overlap is reserved between adjacent areas so that there are enough matching feature points for subsequent image stitching. During the image acquisition process, the system automatically adjusts the exposure time and gain value to ensure high-quality images under different lighting conditions.

[0041] Furthermore, a multi-camera collaborative working mode is introduced, where multiple cameras simultaneously capture images of different areas, further improving the shooting speed and efficiency. Higher-resolution cameras (such as 4K or 8K) are used to obtain finer image details, which is suitable for application scenarios with higher precision requirements.

[0042] Use mathematical operations such as affine transformation or perspective transformation to uniformly transform the coordinates of the measurement points in multiple images. For each overlapping area, select at least 3 fixed feature points as the calibration reference to ensure accurate alignment of the coordinate systems between different images. The algorithm is based on SIFT (Scale Invariant Feature Transform) or ORB (Oriented FAST and RotatedBRIEF) feature detection methods to automatically identify feature points in the image and calculate their relative position relationship. During the coordinate transformation process, the system automatically compensates for geometric deformation caused by factors such as lens distortion and perspective change to ensure the accuracy of the final result.

[0043] Working principle: The system first extracts feature points from each image and uses a matching algorithm to find common feature points between adjacent images. Then, these feature points are used to calculate the geometric transformation relationship between the two images (such as rotation, scaling, translation, etc.).

[0044] Through multiple iterations of optimization, the system gradually adjusts the coordinate system of each image, and finally unifies the measurement points in all images into a common coordinate system, achieving high-precision cross-image size measurement.

[0045] The system first detects multiple feature points (such as corners, edges, etc.) in each image, and uses a matching algorithm to find common feature points between adjacent images. Then, the centroid positions of these feature points are calculated as reference points.

[0046] By aligning the reference points in all images, the system can ensure that the coordinate systems between different images are completely consistent, thereby achieving high-precision cross-image dimension measurement.

[0047] Furthermore, an adaptive threshold algorithm is introduced to dynamically adjust the detection sensitivity of feature points to ensure that feature points can be accurately extracted under different lighting conditions.

[0048] A multi-scale feature detection method is adopted to combine the feature point information at different scales to further improve the robustness and accuracy of the reference points.

[0049] The invention includes an adaptive correction algorithm based on deep learning, which can automatically identify feature points in images acquired at different viewing angles and adjust the images to the same coordinate system based on these feature points. The algorithm processing time does not exceed 50 milliseconds and is suitable for measuring objects with complex geometric shapes.

[0050] The algorithm is based on convolutional neural networks (CNN) and recurrent neural networks (RNN), which can automatically learn the distribution patterns of feature points in images and perform geometric transformation corrections based on these feature points.

[0051] How it works: The system first extracts feature points from each image and predicts the corresponding positions of these feature points in other images through a deep learning model. Then, the geometric transformation matrix between images is calculated using these correspondences to transform all images into the same coordinate system. The adaptive correction algorithm can automatically adjust parameters to adapt to different shooting angles and object shapes, ensuring high-precision coordinate transformation in a variety of complex situations.

[0052] High-speed measurement capability: The system can complete 120-dimension measurement tasks within 1 second, and the average measurement time for each dimension is 8.33 milliseconds. This high-speed measurement capability is particularly suitable for real-time detection on industrial automation production lines, significantly improving production efficiency. The system uses multi-threaded parallel processing technology, and multiple measurement tasks can be performed simultaneously, further improving the measurement speed.

[0053] During the measurement process, the system automatically records the measurement results of each dimension and generates a detailed report for subsequent analysis and quality control.

[0054] Working principle: The system can complete the acquisition and processing of a large number of images in a very short time through the coordinated work of high-speed cameras and servo motors. After each shot, the system will immediately perform image analysis and dimension measurement to ensure the real-time measurement results.

[0055] To further improve the measurement speed, the system uses hardware acceleration technology (such as FPGA and GPU), which significantly shortens the time of image processing and data analysis.

[0056] High magnification image area decomposition: The entire measurement area is broken down into multiple sub-areas, and the image magnification of each sub-area can be set to 2-5 times the original magnification for high-precision measurement. For example, for tiny parts that require extremely high precision, the image magnification can be set to 5 times to capture subtle size differences. The system supports dynamic adjustment of image magnification, and users can select different magnification modes according to specific needs to flexibly respond to different types of measurement tasks.

[0057] Working principle: The system adjusts the focal length of the camera and the lens magnification to achieve high-magnification shooting of different sub-areas. Each time you shoot, the system automatically adjusts the camera parameters to ensure that the image clarity and resolution of each sub-area are optimal. By stitching together images of multiple sub-areas, the system can generate a complete high-resolution image, which is suitable for application scenarios with extremely high precision requirements.

[0058] The present invention also introduces an autofocus function to ensure that the image of each sub-area has the best clarity, further improving the measurement accuracy. The multi-sensor fusion technology is used to combine different types of sensors (such as lasers, ultrasonic waves, etc.) to obtain data in more dimensions, further improving the accuracy and reliability of the measurement.

[0059] Users can adjust the measurement range and accuracy through the software interface without adding additional hardware costs. For example, users can choose low-precision mode (±0.1 mm) or high-precision mode (±0.01 mm), and adjust the measurement speed as needed. The system supports multiple measurement modes (such as single measurement, continuous measurement, batch measurement, etc.), and users can choose the appropriate measurement method according to specific needs. The system also provides a wealth of configuration options, and users can customize measurement parameters (such as exposure time, gain value, lens magnification, etc.) to adapt to different application scenarios.

[0060] Working principle: The system provides an intuitive operation interface through the software interface. Users can quickly adjust the measurement parameters by simply clicking and dragging. All configuration parameters can be saved as templates for easy use next time.

[0061] The system also supports remote control and data transmission functions. Users can remotely monitor and manage measurement equipment through the network to achieve real-time data sharing and collaboration.

[0062] The system integrates an autofocus module with a focus range of 100 mm to infinity and a focus speed of no more than 30 milliseconds. When the distance to the measured object changes, the autofocus function quickly adjusts the lens focus to ensure the clarity of the captured image. The focus module uses phase detection autofocus (PDAF) technology, which can complete the focus operation in a very short time, suitable for shooting needs in high-speed motion scenes.

[0063] Working principle: The system uses a built-in focus sensor to monitor the distance between the lens and the object in real time, and automatically adjusts the focus of the lens based on the distance information. Before each shot, the system automatically performs a focus operation to ensure image clarity. During the focusing process, the system also optimizes based on the surface features of the object (such as texture, color, etc.) to ensure the best focus effect on different materials and surfaces.

[0064] The built-in ambient light sensor can monitor the surrounding light conditions in real time and automatically adjust the camera's exposure time and gain value. The system can maintain stable imaging quality within the lighting range of 0-1000 lux to avoid the impact of light changes on the measurement results. The system is also equipped with an adjustable fill light, and users can choose different light intensities and color temperatures according to actual needs to ensure high-quality images in different environments.

[0065] Working principle: The ambient light sensor monitors the surrounding light intensity in real time and feeds the data back to the control system. The control system automatically adjusts the camera's exposure time and gain value according to the light intensity to ensure that the brightness and contrast of the image remain at the optimal state.

[0066] The fill light can automatically adjust the lighting angle and intensity according to the color and material of the measured object, ensuring clear images on surfaces of different materials and colors.

[0067] Optimization solution: Introduce intelligent lighting system to automatically adjust the parameters of fill light according to the characteristics of the measured object (such as color, reflectivity, etc.) to further improve the imaging quality. Use multi-light source combination technology, combined with different types of light sources (such as LED, halogen lamp, etc.) to obtain more dimensional image information, suitable for complex measurement scenarios.

[0068] The system is equipped with an automatic defect detection module based on convolutional neural network (CNN), which can analyze an image within 0.5 seconds, identify and mark the product parts that do not meet the quality standards. The detection accuracy of this module reaches more than 98%, greatly improving the level of product quality control.

[0069] The defect detection module supports multiple defect types (such as scratches, cracks, stains, etc.), and can customize detection rules and thresholds according to user needs.

[0070] Working principle: The system analyzes the image data through deep learning algorithms, automatically identifies defective areas in the image, and classifies and annotates them according to preset rules. After each inspection, the system generates a detailed report indicating the location, type and severity of the defect.

[0071] The defect detection module also supports integration with the manual review system, and the test results can be sent directly to quality inspectors for review to ensure that product quality meets standards.

[0072] In specific applications, the position of the lens is precisely controlled by a high-precision servo motor to ensure that the lens can be accurately positioned at the predetermined position each time a shot is taken. A high-precision linear guide system is used to move the product to the best shooting position in front of the lens, and to keep it horizontal and parallel to reduce measurement errors caused by angle deviations. A high-definition, high-speed camera mounted on a slide is used to capture product images in different areas, and overlapping areas are retained between adjacent shooting areas to ensure image continuity and integrity and support subsequent stitching. Using a coordinate transformation algorithm, the measurement points in multiple images acquired from different perspectives are unified into a common coordinate system, thereby achieving cross-image dimensional measurement. Fixed features are selected in the overlapping area for calibration, and mathematical operations such as affine transformation or perspective transformation are used to ensure accurate alignment of the coordinate systems between different images to improve measurement accuracy.

[0073] The adaptive correction algorithm based on deep learning can automatically identify and adjust images from different perspectives to the same coordinate system, which is suitable for measuring objects with complex geometric shapes, and the processing time does not exceed 50 milliseconds. The system can complete the measurement tasks of more than 100 dimensions within 1 second, and the average measurement time for each dimension is 8.33 milliseconds, which is particularly suitable for real-time detection needs on industrial automation production lines. Hardware acceleration technology and multi-threaded parallel processing further shorten the time of image processing and data analysis.

[0074] The entire measurement area is divided into multiple sub-areas, and the image magnification of each sub-area can be dynamically adjusted to capture the subtle size differences of tiny parts and meet the strict requirements of the precision manufacturing industry. The autofocus function automatically adjusts the lens focus according to the distance of the measured object to ensure the clarity of the captured image; the ambient light compensation mechanism maintains stable imaging quality under different lighting conditions. Equipped with an automatic defect detection module based on deep learning, it can quickly identify and mark product parts that do not meet quality standards, assist manual review, and improve product quality control.

[0075] In this document, the directional words such as front, back, top, and bottom are defined by the positions of the components in the drawings and the positions of the components relative to each other, and are only for the sake of clarity and convenience in expressing the technical solution. It should be understood that the use of the directional words should not limit the scope of protection claimed in this application.

[0076] Obviously, the embodiments described above are only part of the embodiments of the present application, rather than all of the embodiments. The present application is not limited to the details of the above embodiments, and any appropriate changes or modifications made by ordinary technicians in the relevant technical field are deemed to be within the patent scope of the present application.

Claims

1. A method for measuring cross-image dimensions by coordinate transformation, characterized in that: include: Use servo motors to control the timing and position of lens movement to ensure that the lens is positioned correctly for each shot; The product is moved to a suitable position through the feeding mechanism to ensure that the measuring object is horizontal and parallel to the lens, thereby reducing the measurement error caused by angle deviation; Use high-definition high-speed cameras to capture images in different areas along a predetermined path on the slide rail, and retain a certain proportion of overlapping areas between adjacent shooting areas to ensure image continuity and integrity, and support subsequent image stitching; A coordinate transformation algorithm is used to unify the measurement points in multiple images into a common coordinate system.

2. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 1, characterized in that: The fixed features in the overlapping area of ​​adjacent images are further used for calibration, and mathematical operations are used to ensure the accurate alignment of the coordinate systems between different images.

3. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 1, characterized in that: The center coordinates of the features in the overlapping area of ​​each image are calculated and used as the reference point for the uniform alignment of the coordinates of all measurement points.

4. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 1, characterized in that: During the coordinate conversion process, an adaptive correction algorithm is used to accurately convert images acquired at different viewing angles into the same coordinate system.

5. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 1, characterized in that: This method can complete the measurement of more than 100 dimensions within 1 second, while providing measurement speed and accuracy far exceeding existing technologies, and can be used for real-time detection on industrial automation production lines.

6. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 1, characterized in that: The measurement area is decomposed into multiple image areas, and the image magnification is increased several times for high-precision measurement.

7. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 5, characterized in that: This method can greatly improve the flexibility and adaptability of the measurement system without increasing additional hardware costs, allowing users to adjust the measurement range and accuracy according to specific needs.

8. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 1, characterized in that: Also included is an autofocus function that automatically adjusts the lens focus according to the distance of the measured object, ensuring that images taken at different positions have optimal clarity.

9. The method for realizing cross-image dimension measurement by coordinate transformation according to claim 1, characterized in that: This method is equipped with an ambient light compensation mechanism to maintain stable imaging quality under different lighting conditions and avoid measurement errors caused by light changes.

10. The method for realizing cross-image dimension measurement through coordinate transformation according to claim 1, characterized in that: It also includes built-in automatic defect detection capabilities that analyze image data based on deep learning algorithms to quickly identify and mark product parts that do not meet quality standards.

Images