Visual identification and positioning method and system for precast concrete member mold

Abstract

The invention belongs to the technical field of concrete prefabrication, and relates to a prefabricated concrete member mold visual identification and positioning method and system.The method comprises the steps that when a mold is conveyed to a designated position through a track, the mold is initially positioned; under the condition that the initial positioning of the mold is within a preset range, performing image acquisition on the mold; the collected mold image is preprocessed; analyzing the preprocessed mold image, and determining a mold type and a positioning mark; determining the reference position of the mold according to the mold type; the deviation distance and angle between the positioning mark of the mold and the reference position are calculated, and the actual position offset of the mold is obtained; and the mold is machined based on the actual position offset of the mold. The device has the beneficial effects that the automation degree of a production line is remarkably improved, manual intervention is reduced, and the production cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of concrete prefabrication, and in particular to a method and system for visually identifying and positioning a prefabricated concrete component mold. Background Art

[0002] Precast concrete component molds play a key role in concrete prefabrication. They mainly include double-block sleeper molds, prestressed sleeper molds, shield segment molds, PC laminated plate molds, and railway assembled cable trough molds. At the same time, in the production of these molds and the necessary links such as mold cleaning, the positioning of the mold is the key. If the mold cannot be accurately identified and positioned, the subsequent processing of the mold cannot be automated, and the production efficiency cannot be improved.

[0003] Traditional mold positioning methods mainly rely on manual operation or simple mechanical positioning devices, which have significant limitations in modern production. First, the positioning accuracy is low. Manual operation is affected by the skills, experience and even physical condition of the operator, and it is difficult to ensure high accuracy and consistency of positioning. Simple mechanical positioning devices are limited by the limitations of the mechanical structure itself and are also difficult to achieve high-precision positioning. Secondly, the adaptability is poor. In the face of different types of precast concrete component molds, traditional methods often need to adjust or replace dedicated positioning devices, which limits production flexibility and increases production costs and time; thirdly, the degree of automation is low. Traditional methods lack compatibility with modern automated production lines and cannot fully utilize the advantages brought by automation technology, limiting the level of automation and intelligence in the production process; finally, process docking is difficult. Due to the limitations of positioning accuracy and efficiency, additional adjustment and docking work is required when the mold positioning results are applied to subsequent processes, which increases the complexity of the work and the error rate. Summary of the invention

[0004] Technical issues to be solved In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a method and system for visual identification and positioning of precast concrete component molds, which solves the technical problem of low automation and intelligence levels of traditional mold positioning methods.

[0005] Technical Solution In order to achieve the above object, the main technical solutions adopted by the present invention include: In a first aspect, the present invention provides a method for visually identifying and positioning a mold for a precast concrete component, comprising: When the mold is sent to the designated position through the track, the mold is initially positioned; When the initial positioning of the mold is within a preset range, an image of the mold is captured; Preprocessing the collected mold images; Analyze the pre-processed mold image to determine the mold type and positioning mark; Determine the mold reference position according to the mold type; Calculate the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold; The mold is processed based on the actual position offset of the mold.

[0006] Optionally, the collected mold image is preprocessed, including: The collected mold images are subjected to denoising, contrast enhancement and histogram equalization processing.

[0007] Optionally, the pre-processed mold image is analyzed to determine the mold type, including: The preprocessed mold image is analyzed based on template matching or deep learning recognition model to determine the mold type.

[0008] Optionally, analyzing the preprocessed mold image based on template matching to determine the mold type includes: Collect images of all molds to be identified in advance, select feature areas on the images according to different mold types, and use them as feature image templates for the corresponding molds; The feature image of the mold to be classified is matched with the feature root image template by sliding window, and the mold category of the current feature image is determined according to the matching threshold.

[0009] Optionally, the preprocessed mold image is analyzed based on a deep learning recognition model to determine the mold type, including: Collect a preset number of mold images for each type of mold according to different lighting, shooting distances, and mold placement angles, and train a deep learning recognition model based on a deep learning algorithm; Use the deep learning recognition model to analyze the preprocessed mold image and determine the mold type.

[0010] Optionally, the deep learning recognition model adopts a VGG-16 network and uses Leaky ReLU as an activation function.

[0011] Optionally, calculating the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold includes: Get the pixel coordinates of the center point of the positioning mark; Compare the coordinates with the pixel coordinates of the center point of the reference position; Calculate the center point offset pixel of the mold positioning mark; Compare the center line of the positioning mark with the center line of the reference position; Calculate the offset angle of the mold positioning mark; The ratio of the pixel to the actual distance is calculated based on the pixel width of the positioning mark and the actual width of the mold positioning mark, and converted into the actual position offset of the mold.

[0012] In a second aspect, the present invention provides a precast concrete component mold visual recognition and positioning system, comprising: The mold initial positioning module performs initial positioning on the mold when the mold is sent to the designated position through the track; An image acquisition module is used to acquire images of the mold when the initial positioning of the mold is within a preset range; An image preprocessing module preprocesses the collected mold images; The visual recognition module analyzes the pre-processed mold image to determine the mold type and positioning mark; and determines the mold reference position according to the mold type; An offset calculation module calculates the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold; The system integration and automation application module processes the mold based on the actual position offset of the mold.

[0013] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the method for visual recognition and positioning of a precast concrete component mold as described in any one of the first aspects above.

[0014] In a fourth aspect, the present invention provides a storage device, including a storage medium and a processor, wherein the storage medium stores a computer program, and when the program is executed by the processor, the method for visual recognition and positioning of a precast concrete component mold as described in any one of the first aspects above is implemented.

[0015] Beneficial Effects The beneficial effects of the present invention are as follows: a method for visual identification and positioning of a precast concrete component mold of the present invention is intended to improve the mold positioning method, improve the efficiency of mold positioning, and enhance the automation and intelligence level of the production system. The present invention realizes the rapid and accurate identification and positioning of different types of molds through the steps of initial mold positioning, image acquisition, image preprocessing, visual recognition algorithm, positioning deviation calculation, system integration and automated application. According to the identification and positioning results of the mold, the present invention can automatically calculate the deviation between the mold and the reference position, and can also seamlessly connect with subsequent processes such as automatic polishing, cleaning or spraying, thereby improving the automation level of the production line. The implementation of the present invention will significantly improve the automation level of the production line, reduce manual intervention, and reduce production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a flow chart of a method for visually identifying and positioning a mold for a precast concrete component provided by an embodiment of the present invention; Figure 2 A schematic diagram of initial positioning of a mold provided by an embodiment of the present invention; Figure 3 A schematic diagram of image acquisition provided by an embodiment of the present invention; Figure 4 A flowchart of image preprocessing and visual recognition provided by an embodiment of the present invention; Figure 5 A structural diagram of the VGG-16 model provided in an embodiment of the present invention; Figure 6 A schematic diagram of system integration and automation application provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below through specific implementation modes in conjunction with the accompanying drawings.

[0018] The embodiment of the present invention proposes a method for visually identifying and positioning a precast concrete component mold, which aims to overcome the limitations of traditional mold positioning methods, improve the accuracy and efficiency of mold positioning, and enhance the automation and intelligence level of the production line system. The method is applicable to a variety of precast concrete component molds, including track plate molds, double-block sleeper molds, prestressed sleeper molds, shield segment molds, PC composite plate molds, railway assembled cable trough molds, etc.

[0019] In order to better understand the above technical solution, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a clearer and more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0020] First, refer to Figure 1 This embodiment provides a method for visually identifying and positioning a precast concrete component mold, including: S1, when the mold is sent to the designated position through the track, the mold is initially positioned.

[0021] Combination Figure 2 The mold moves forward along the running direction on the conveyor track, and a sensor sensing area is set just below the center of the mold. The sensing area should be an object that can be detected by the sensor alone. When the mold moves forward to the position where the sensing area is just detected by the sensor, the mold is automatically triggered to stop moving forward, thereby achieving the effect of initial positioning.

[0022] When the mold is sent to the designated position via the track, an initial positioning mechanism is used to ensure that the mold is in the approximately correct position and orientation. This step is critical for subsequent accurate visual recognition and positioning. The initial positioning mechanism can use a simple sensor device to reduce the lateral deviation of the mold on the track. Ensure that the mold position meets the image acquisition requirements.

[0023] S2, when the initial positioning of the mold is within a preset range, an image of the mold is captured.

[0024] Use a high-resolution camera and an appropriate light source combination to capture images of the mold that has entered the designated position and has undergone initial positioning. The camera should have sufficient resolution to capture the details of the mold, and the light source combination should ensure uniform brightness of the image to avoid the impact of shadows and reflections on the image quality. Provide a data basis for subsequent identification and positioning.

[0025] Reference Figure 3 , the camera should be installed above the mold conveyor track to obtain the best viewing angle. The combination of light sources needs to be adjusted according to the actual environment to ensure that the brightness and contrast of the image are optimal. In addition, the camera can choose an autofocus industrial camera to obtain a clear image.

[0026] S3, preprocessing the collected mold image.

[0027] S4, analyzing the pre-processed mold image to determine the mold type and positioning mark.

[0028] S5, determining the mold reference position according to the mold type.

[0029] Combination Figure 3 , the dotted box is the camera field of view, which must cover the characteristic area of ​​the mold, that is, the type of the current mold can be determined through this characteristic area. First, in the center of the camera field of view, at a position perpendicular to the edge of the mold on the ground, install or paste a ground positioning mark, the color characteristics of the ground positioning mark must be clearly different from the ground and the mold body; secondly, install or paste a mold positioning mark on the edge of the mold. The positioning mark is roughly a rectangle perpendicular to the edge of the mold, and the color characteristics are clearly different from the mold body. The position of the positioning mark needs to be aligned with the ground positioning mark as the reference position of the mold positioning mark.

[0030] S6, calculating the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold.

[0031] This step needs to be based on the position of the positioning mark in the recognition result to ensure the accuracy of positioning.

[0032] S7, processing the mold based on the actual position offset of the mold.

[0033] Combination Figure 6 , the system integration is carried out through a data interaction module, which passes the offset data obtained in the above method to the subsequent process. Subsequent processes such as grinding, cleaning, spraying or other automated processes all require corresponding preset process points, and the subsequent processes are automated based on these preset process points and the reference position of the mold positioning mark. After obtaining the offset data passed, the subsequent process attaches the offset data to the preset process point to obtain the actual process point, and then directly performs the corresponding automated processing based on the actual process point to realize automated application.

[0034] In summary, the present invention provides a highly automated and intelligent method for visual identification and positioning of precast concrete component molds, which realizes mold identification and positioning through image processing and machine vision technology, and can improve the degree of automation of subsequent processes of the production line. At the same time, the adaptability and flexibility of the production line have also been significantly improved. This method is expected to be widely used in the production process of precast concrete components in the field of rail transit construction, promoting technological progress and industrial upgrading in the industry.

[0035] Optionally, the collected mold image is preprocessed, including: The collected mold images are subjected to denoising, contrast enhancement and histogram equalization processing.

[0036] These operations can improve image quality and recognition accuracy, laying a good foundation for subsequent visual recognition algorithms.

[0037] Optionally, the pre-processed mold image is analyzed to determine the mold type, including: The preprocessed mold image is analyzed based on template matching or deep learning recognition model to determine the mold type.

[0038] This step is the key to achieving high-precision positioning, and requires training and optimizing algorithm models to improve recognition accuracy.

[0039] Optionally, analyzing the preprocessed mold image based on template matching to determine the mold type includes: Collect images of all molds to be identified in advance, select feature areas on the images according to different mold types, and use them as feature image templates for the corresponding molds; The feature image of the mold to be classified is matched with the feature root image template by sliding window, and the mold category of the current feature image is determined according to the matching threshold.

[0040] Optionally, the preprocessed mold image is analyzed based on a deep learning recognition model to determine the mold type, including: Collect a preset number of mold images for each type of mold according to different lighting, shooting distances, and mold placement angles, and train a deep learning recognition model based on a deep learning algorithm; Use the deep learning recognition model to analyze the preprocessed mold image and determine the mold type.

[0041] Optionally, the deep learning recognition model adopts a VGG-16 network and uses Leaky ReLU as an activation function.

[0042] The present invention adopts an improved VGG-16 network to train the above classification model, and uses Leaky ReLU as the activation function instead of the original ReLU activation function, which optimizes the problem that the excessive sparsity of ReLU units in the original VGG-16 network causes a large number of neurons to fail to activate. The improved VGG-16 network structure is as follows Figure 5 As shown in the figure, it contains 13 convolutional layers + Leaky ReLU activation function and 3 fully connected layers + Leaky ReLU activation function. The pooling layer and softmax output layer remain unchanged. The formula of the Leaky ReLU activation function is as follows:

[0043] Among them, α can expand the range of the ReLU function, and usually takes a small positive number as a non-zero slope, such as 0.01. This slope allows the output to no longer be 0 when the input is negative, thereby retaining negative value information.

[0044] Optionally, calculating the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold includes: Get the pixel coordinates of the center point of the positioning mark; Compare the coordinates with the pixel coordinates of the center point of the reference position; Calculate the center point offset pixel of the mold positioning mark; Compare the center line of the positioning mark with the center line of the reference position; Calculate the offset angle of the mold positioning mark; The ratio of the pixel to the actual distance is calculated based on the pixel width of the positioning mark and the actual width of the mold positioning mark, and converted into the actual position offset of the mold.

[0045] See also Figure 4 , is a flowchart of image preprocessing and visual recognition in the present invention.

[0046] After the mold type is identified based on template matching or deep learning recognition model, if the mold type still cannot be identified, the original image is preprocessed again using the next set of image preprocessing parameters, and then the mold type is identified; if there is no next set of preprocessing parameters, it means that the identification has failed and the process ends.

[0047] After the mold type is identified based on template matching or deep learning recognition model, if the mold type has been identified, template matching is used to find the positioning mark. The positioning mark includes Figure 3 The mold positioning mark and the ground positioning mark shown in the figure, the template matching is to use the images of the two mold positioning marks as templates, and perform sliding window matching on the current image to be recognized. If one of the positioning marks cannot be found, the recognition fails and the process ends. If both positioning marks have been found, the next step of the process is entered.

[0048] Calculate the offset of the positioning mark, that is, calculate the offset between the mold positioning mark and its reference position through the relative position of the two positioned positioning marks, and convert it into the offset of the actual position of the mold. The completion of the conversion indicates that the recognition is successful and the process ends.

[0049] The method for visually identifying and positioning a precast concrete component mold provided in this embodiment has the following advantages: 1. High-precision positioning: Through image processing and visual algorithms, high-precision recognition and positioning of the mold are achieved.

[0050] 2. High-efficiency production: Automated visual recognition and positioning greatly improve the efficiency of the production line, reduce manual intervention, and shorten the production cycle.

[0051] 3. Strong adaptability: It can automatically identify and locate different types of molds without replacing or adjusting special positioning devices, which improves the flexibility and adaptability of the production system.

[0052] 4. High degree of automation and intelligence: The present invention is compatible with modern automated production lines and can automatically connect to subsequent processes, thus improving the automation and intelligence level of the overall production.

[0053] 5. Reduce production costs: Reduce the scrap rate caused by inaccurate positioning, reduce production costs, and improve product consistency and quality.

[0054] In the second aspect, the present embodiment provides a precast concrete component mold visual recognition and positioning system, including: a mold initial positioning module, when the mold is sent to the specified position through the track, the mold is initially positioned; an image acquisition module, when the initial positioning of the mold is within the preset range, the mold is imaged; an image preprocessing module, preprocessing the acquired mold image; a visual recognition module, analyzing the preprocessed mold image to determine the mold type and positioning mark; determining the mold reference position according to the mold type; an offset calculation module, calculating the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold; a system integration and automation application module, processing the mold based on the actual position offset of the mold. According to the precast concrete component mold visual recognition and positioning system provided by this embodiment, since it is used to implement the steps of a precast concrete component mold visual recognition and positioning method provided by the first aspect of the present invention, the precast concrete component mold visual recognition and positioning system has all the technical effects of the precast concrete component mold visual recognition and positioning method, which will not be repeated here.

[0055] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements a method for visually identifying and positioning a mold for a precast concrete component as described in any one of the first aspects above.

[0056] In a fourth aspect, an embodiment of the present invention provides a storage device, including a storage medium and a processor, wherein the storage medium stores a computer program, and when the program is executed by the processor, the method for visual recognition and positioning of a precast concrete component mold as described in any one of the first aspects above is implemented.

[0057] It should be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0058] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention should also include these modifications and variations.

[0059] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may alter, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A method for visually identifying and positioning a precast concrete component mold, characterized in that: include: When the mold is sent to the designated position through the track, the mold is initially positioned; When the initial positioning of the mold is within a preset range, an image of the mold is captured; Preprocessing the collected mold images; Analyze the pre-processed mold image to determine the mold type and positioning mark; Determine the mold reference position according to the mold type; Calculate the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold; The mold is processed based on the actual position offset of the mold.

2. A method for visually identifying and positioning a precast concrete component mold according to claim 1, characterized in that: Preprocess the collected mold images, including: The collected mold images are subjected to denoising, contrast enhancement and histogram equalization processing.

3. A method for visual identification and positioning of a precast concrete component mold according to claim 2, characterized in that: Analyze the pre-processed mold image to determine the mold type, including: The preprocessed mold image is analyzed based on template matching or deep learning recognition model to determine the mold type.

4. A method for visually identifying and positioning a precast concrete component mold according to claim 3, characterized in that: Analyze the preprocessed mold image based on template matching to determine the mold type, including: Collect images of all molds to be identified in advance, select feature areas on the images according to different mold types, and use them as feature image templates for the corresponding molds; The feature image of the mold to be classified is matched with the feature root image template by sliding window, and the mold category of the current feature image is determined according to the matching threshold.

5. A method for visually identifying and positioning a precast concrete component mold according to claim 4, characterized in that: Analyze the preprocessed mold image based on the deep learning recognition model to determine the mold type, including: Collect a preset number of mold images for each type of mold according to different lighting, shooting distances, and mold placement angles, and train a deep learning recognition model based on a deep learning algorithm; Use the deep learning recognition model to analyze the preprocessed mold image and determine the mold type.

6. A method for visually identifying and positioning a mold for a precast concrete component according to claim 5, characterized in that: The deep learning recognition model adopts the VGG-16 network and uses Leaky ReLU as the activation function.

7. A method for visually identifying and positioning a precast concrete component mold according to claim 6, characterized in that: Calculate the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold, including: Get the pixel coordinates of the center point of the positioning mark; Compare the coordinates with the pixel coordinates of the center point of the reference position; Calculate the center point offset pixel of the mold positioning mark; Compare the center line of the positioning mark with the center line of the reference position; Calculate the offset angle of the mold positioning mark; The ratio of the pixel to the actual distance is calculated based on the pixel width of the positioning mark and the actual width of the mold positioning mark, and converted into the actual position offset of the mold.

8. A visual recognition and positioning system for precast concrete component molds, characterized in that: include: The mold initial positioning module performs initial positioning on the mold when the mold is sent to the designated position through the track; An image acquisition module is used to acquire images of the mold when the initial positioning of the mold is within a preset range; An image preprocessing module is used to preprocess the collected mold images; The visual recognition module analyzes the pre-processed mold image to determine the mold type and positioning mark; Determine the mold reference position according to the mold type; An offset calculation module calculates the deviation distance and angle between the positioning mark of the mold and the reference position to obtain the actual position offset of the mold; The system integration and automation application module processes the mold based on the actual position offset of the mold.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for visually identifying and positioning a precast concrete component mold as claimed in any one of claims 1 to 7 is implemented.

10. A storage device comprising a storage medium and a processor, wherein the storage medium stores a computer program, characterized in that: When the processor executes the computer program, the method for visually identifying and positioning a precast concrete component mold as claimed in any one of claims 1 to 7 is implemented.

Images