Double-track motion control method

Abstract

The invention provides a double-track motion control method, and relates to the technical field of visual detection control. The method comprises the steps of hardware configuration, starting operation, position triggering, image processing, result processing and sorting. A double-rail conveying system is adopted in hardware, and a high-precision DD motor, a high-transmittance glass disc, an adjustable light source and an upper and lower detection camera are arranged; after starting, the DD motor rotates at a constant speed, and the vibration disc feeds; when the position is triggered, the encoder is combined to feed back the dynamic compensation detection opportunity; image processing adopts a multi-thread / multi-process parallel framework and introduces priority scheduling; and finally, visual results are combined through custom scripts, and blowing sorting is triggered. According to the invention, the problems of double-track synchronous collaboration and multi-source data real-time processing are solved, the detection efficiency is improved by more than 1.8 times, the space utilization rate of equipment is improved, complex sorting logic is realized, the detection stability and accuracy are guaranteed, the adaptability is enhanced, and the method is suitable for high-requirement visual detection scenes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of visual detection control, in particular to a double-track motion control method. BACKGROUND

[0002] In the field of visual detection, with the continuous improvement of the detection efficiency and accuracy requirements of industrial production, the traditional single-track visual detection system has gradually been difficult to meet the actual production needs. The single-track visual detection system usually adopts a single track to convey workpieces, processes image data through a single thread or a single GPU core, and can only perform simple visual and blowing port binding to complete sorting tasks.

[0003] As the known technology basis of the present application, the visual detection system is generally based on CCD imaging and image processing technology, adopts a CCD camera to convert the detected target into an image signal, converts it into a digital signal according to pixel distribution, brightness, color and other information, and an image processing system operates on these signals to extract target features, and then outputs the results according to the preset conditions to realize automatic detection and recognition. The detection and classification process is usually that a vibrating feeding mechanism arranges and conveys products to a glass turntable, a detection module beside the glass turntable module takes pictures, analyzes the data, and then the products enter a sorting module for classification.

[0004] The closest technical solution to the present application is a single-track visual detection control method, which has the following steps: a single-track conveying system is adopted, a single glass disc and corresponding detection cameras, light sources and the like are provided; a driving motor is started to make the glass disc rotate, and a vibrating disc feeder is turned on; when the product passes through the set camera position, the camera is triggered to take a picture; the camera transmits the image to the vision system, the vision system processes the image data using a single thread or a single GPU core, and completes the detection task; according to the detection result, the vision result is bound with a specific blowing port, and when the material reaches the corresponding blowing position, the blowing sorting is triggered.

[0005] The existing single-track visual detection control method has the following disadvantages: (1) low detection efficiency: single-track conveying and single-thread / single-GPU core processing of image data limit the number of workpieces detected per unit time, and the traditional single-track visual detection speed is only 1000-1500 pieces per minute, which cannot meet the high-yield production demand; (2) low space utilization of equipment: the single-track structure can only utilize part of the space of the equipment, and cannot fully develop the maximum potential of the equipment; (3) simple sorting logic: only simple visual and blowing port binding can be performed, fine result combination of visual tasks cannot be performed, and it is difficult to complete complex sorting logic; (4) poor adaptability: it is difficult to adjust and adapt to workpieces of different types and different detection requirements, and cannot flexibly meet diversified detection needs. SUMMARY

[0006] In view of the problems, the present application is proposed to provide a double-track motion control method to overcome the problems or at least partially solve the problems, aiming to solve the problems of double-track synchronization and coordination control in a double-track glass disc visual inspection system and multi-source data parallel processing and real-time performance.

[0007] In some embodiments of the present application, a double-track motion control method is disclosed, 1. A double-track motion control method, characterized by comprising the following steps: A double-track conveying system is used, the distance between the two tracks is 9mm, the track surface is covered with a hard oxide layer, the glass disc is made of super white high transmittance glass with a transmittance of >92% and the disc flatness error is controlled within 0.04mm, the driving system uses DD motor to ensure that the glass disc rotation speed is stable at 30-60rpm; The upper and lower detection cameras are configured for focusing on the two ends of the glass tube to detect defects, point light is used to illuminate the two ends, and the color temperature of the light source is adjusted to 3000K-6500K according to the color of the embroidery texture; Start the equipment, make the DD motor rotate to uniform speed, and open the vibration disc; Determine the product origin through the optical fiber sensor, when the product passes through the set camera position, trigger the camera to take pictures combined with the high-speed position comparison result, at the same time, real-time acquisition of the encoder feedback signal of the double-track, dynamic calculation of the position difference and speed difference of the two tracks, compensation of the detection trigger timing; The camera transmits the collected images to the vision system in real time, the vision system uses a multi-thread / multi-process parallel processing framework with pipeline splitting from image acquisition, preprocessing, detection to final result output, allocates independent computing resources to the double-track, introduces a priority scheduling algorithm, temporarily increases the resource proportion when a complex defect occurs in a track, and processes complex vision tasks; Script processing of the vision detection results is performed in real time, the air blowing port is planned, the air blowing action is triggered when the corresponding material reaches the planned air blowing position, and the sorting is completed; The loading process can choose double-track loading, single-track loading or double-supply mode.

[0008] Optionally, the way of compensating the detection trigger timing is adjusting the camera exposure time, trigger delay or correcting the ROI position of the detection area.

[0009] Optionally, the independent computing resources include CPU multi-core and GPU multi-stream processor.

[0010] Optionally, the script processing of the vision detection results is performed in real time, the air blowing port is planned, the air blowing action is triggered when the corresponding material reaches the planned air blowing position, and the sorting is completed; It includes: The visual detection result is processed in real time by using a self-defined script, a blowing port is planned, and when the corresponding material reaches the planned blowing position, the blowing action is triggered by comprehensively controlling the blowing pulse position, and the sorting is completed.

[0011] Optionally, the resolution of the upper and lower detection cameras is not less than 1.3 million pixels, the frame rate is not less than 200 fps, and the lens focal length is 12-50 mm.

[0012] Optionally, the response time of the optical fiber sensor is not more than 70 s, and the detection position error is ±0.1 mm.

[0013] Optionally, the air pressure of the blowing action is adjusted in the range of 0.2-0.6 MPa, and the blowing time is controlled in the range of 8-12 ms.

[0014] The application has the following advantages: In the embodiment of the application, by using a double-track conveying system and a multi-thread / multi-process parallel processing framework, the double-track visual detection system is 1.8 times faster than the single-track detection system, can meet the requirements of customers for higher detection speed (such as more than 2500 / min), and effectively solves the bottleneck of factory efficiency improvement; the double-track design can utilize the synchronous operation of the inner and outer tracks of the optical screening machine glass disc or half operation, and does not interfere with each other, fully utilizes the maximum space of the equipment, realizes the synchronous detection of the product double channel, or the simultaneous detection of different products; through real-time combination of the visual results by using self-defined script processing, complex sorting tasks can be completed, and the limitation of traditional single-track that can only simply bind vision and blowing port is broken; on the hardware, high-precision configuration is adopted, such as DD motor to ensure stable speed, high-transmittance and flatness glass disc to avoid image distortion, and adjustable light source to ensure image contrast; on the software, real-time position compensation and priority scheduling algorithm are used to ensure double-track synchronous cooperation and real-time data processing, improve detection stability and accuracy; supports double-track feeding, single-track feeding, double-supply and other feeding modes, can be adjusted according to different detection requirements, can meet diversified detection scenes, and can be non-standard customized to meet different detection requirements of users. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the application, the following will briefly introduce the drawings needed to be used in the description of the application. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creating any creative labor.

[0016] Fig. 1 is a step flow chart of a double-track motion control method provided by an embodiment of the application; Fig. 2Fig. 1 is a schematic diagram of a logic structure of a double-track motion control method according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to make the objectives, characteristics and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0018] Reference Figs. 1-2 An embodiment of the present application provides a double-track motion control method, and specific technical solutions are as follows. Step S1, hardware configuration: a double-track conveying system is adopted, the distance between the two tracks is set to 9 mm, the track surface is covered with a hard oxide layer to reduce frictional damage when the glass tube is conveyed; the glass disc is made of super-white high-transmittance glass with a light transmittance of >92%, and the disc flatness error is controlled within 0.04 mm to avoid image distortion caused by disc deformation; a DD motor is adopted for the driving system to ensure that the glass disc rotation speed is stably controlled at 30-60 rpm, so that the conveying beat and the detection rhythm are accurately matched. The upper and lower detection cameras are configured to focus on both ends of the glass tube, and the end defects such as cracking and burrs are detected, and a point light source is used to illuminate both ends; according to the embroidery texture color, the color temperature of the light source is adjusted to 3000K-6500K to ensure that the texture image contrast is stable.

[0019] Step S2, start running: after starting the equipment, when the DD motor rotates to a uniform speed state, the vibrating disc is turned on to start feeding, and the feeding process can select double-track feeding, single-track feeding or double-feeding mode.

[0020] Step S3, position triggering: the product origin is determined by the optical fiber sensor, and when the product passes through the set camera position, the camera is triggered to take a picture in combination with the high-speed position comparison result. At the same time, the encoder feedback signal (such as pulse signal) of the double track is obtained in real time, the position difference and speed difference of the two tracks are dynamically calculated, and the detection triggering time is compensated by adjusting the camera exposure time, trigger delay or correcting the ROI position of the detection area, so as to ensure the spatial alignment of the workpieces on both sides in the detection area.

[0021] Step S4, image processing: the camera transmits the collected images to the vision system in real time. The vision system adopts a multi-thread / multi-process parallel processing framework that splits the pipeline from image acquisition to preprocessing, detection, and finally result output. Independent CPU multi-core, GPU multi-stream processor, and other computing resources are allocated to the two tracks to avoid data blocking on one track affecting the processing of the other track. A priority scheduling algorithm is introduced. When a complex defect appears on one track (which requires longer computation time), the resource allocation ratio of that track is temporarily increased to ensure that critical detection tasks are not delayed and complex vision tasks are processed efficiently.

[0022] Step S5, result processing and sorting: the vision detection results are processed and combined in real time using a custom script. The air blowing port is planned, and when the corresponding material reaches the planned air blowing position, the air blowing action is triggered by controlling the air blowing pulse position, and the sorting is completed.

[0023] The dual-track conveying system and multi-thread / multi-process parallel processing framework are adopted. The dual-track vision detection system has a speed improvement of 1.8 times compared to the single-track detection system, which can meet the customer's requirement for higher detection speed (e.g., greater than 2500 per minute) and effectively solve the bottleneck of factory efficiency improvement. The dual-track design can utilize the inner and outer tracks of the optical screening machine glass disc to run synchronously or run in half, without interfering with each other, fully utilizing the maximum space of the equipment, realizing product double-channel synchronous detection, or simultaneous detection of different products. Through custom script processing of the vision results, complex sorting tasks can be completed, breaking through the limitations of traditional single-track simple binding of vision and air blowing port. On the hardware side, high-precision configurations are adopted, such as DD motor to ensure stable speed, high-transparency and flatness glass disc to avoid image distortion, and adjustable light source to ensure image contrast. On the software side, real-time position compensation and priority scheduling algorithm are used to ensure dual-track synchronization and real-time data processing, improving detection stability and accuracy. It supports dual-track feeding, single-track feeding, and dual-supply feeding, and can be adjusted according to different detection requirements to meet various detection scenarios and non-standard customization to meet different detection needs of users.

[0024] In the following, a dual-track motion control method in the present exemplary embodiment will be further described.

[0025] As an example, 1. A dual-track motion control method, characterized by comprising the following steps: Step S1, a double-track conveying system is adopted, the distance between the two tracks is 9 mm, the track surface is covered with a hard oxide layer, the glass disc is selected from super white high light transmission glass with a light transmission rate of > 92% and the disc flatness error is controlled within 0.04 mm, the driving system adopts a DD motor to ensure that the rotation speed of the glass disc is stable at 30-60 rpm; the upper and lower detection cameras are configured for focusing on the two end defects of the glass tube, point light is used to illuminate the two ends, and the color temperature of the light source is adjusted to 3000K-6500K according to the color of the embroidery texture; Step S2, start the equipment, and when the DD motor rotates to a uniform speed, open the vibration disc; Step S3, determine the product origin through the optical fiber sensor, when the product passes through the set camera position, trigger the camera to take a picture in combination with the high-speed position comparison result, and at the same time, real-time encoder feedback signals of the double tracks are obtained, the position difference and speed difference of the two tracks are dynamically calculated, and the detection trigger timing is compensated; Step 4, the camera transmits the collected images to the vision system in real time, the vision system adopts a multi-thread / multi-process parallel processing framework with pipeline splitting from image acquisition, preprocessing, detection to final result output, allocates independent computing resources for the double tracks, introduces a priority scheduling algorithm, temporarily increases the resource occupancy ratio when a complex defect occurs in a track, and processes complex vision tasks; Step 5, the vision detection results are processed in real time by scripts, the blowing port is planned, when the corresponding material reaches the planned blowing position, the blowing action is triggered, and the sorting is completed; the feeding process can select double-track feeding, single-track feeding or double-feeding mode.

[0026] The two core technical problems of double-track synchronous cooperation and real-time processing of multi-source data are solved; on the one hand, the performance is improved in multiple dimensions: the detection efficiency is improved by more than 1.8 times, the device space is fully utilized, the complex sorting logic is realized, the detection stability and accuracy are guaranteed, and the multi-dimensional performance improvement of multiple feeding modes is supported, which meets the demand of high requirement detection scene; on the other hand, a complete technical closed loop is formed: from start to sorting, the whole process design ensures the smoothness and effectiveness of the double-track motion control, and provides a complete and reliable technical scheme for practical application.

[0027] As an example, the double-track motion control method of the present application is described in detail by taking the detection of broken gaps of embroidery glass tubes as an example. The embroidery glass tube is usually a cylindrical structure (length 1-3 mm, diameter 0.5-2 mm), and the surface glass is easy to wear due to collision, so the detection accuracy, texture recognition ability and detection efficiency are extremely high.

[0028] The hardware configuration comprises a double-track conveying system: two tracks with a spacing of 9 mm, the track surface is covered with a hard oxide layer to reduce frictional damage during conveying of the embroidered glass tube; the glass disc is made of super-white high-transmittance glass with a light transmittance of more than 92%, and the disc flatness error is controlled within 0.04 mm to prevent image distortion of the embroidered glass tube caused by disc deformation; the driving system adopts a DD motor to accurately control the glass disc rotation speed to be stable at 30-60 rpm, so that the embroidered glass tube conveying rhythm and detection rhythm are accurately matched.

[0029] The visual detection system comprises upper and lower detection cameras configured to focus on the two ends of the embroidered glass tube respectively, for detecting end cracking, burr and other defects, and a point light source is used to illuminate the two ends to ensure clear imaging of the end defects; according to the color of the texture of the embroidered glass tube, the color temperature of the light source is adjusted to 3000K-6500K to ensure stable contrast of the texture image, facilitating subsequent texture recognition and crack gap detection.

[0030] The start-up operation comprises: after starting the device, the DD motor starts to operate and gradually accelerates to a uniform speed of 30-60 rpm, after the speed is stable, the vibrating disc is turned on, the vibrating disc arranges the embroidered glass tubes in order and conveys them to the double-track conveying system, and in this embodiment, a double-track feeding mode is adopted.

[0031] The position triggering step comprises: the embroidered glass tube moves with the double-track conveying system, and when passing through the optical fiber sensor, the original position of the embroidered glass tube is determined. As the double-track continues to convey, when the embroidered glass tube passes through the set camera position, the system triggers the camera to take a picture in combination with the high-speed position comparison result. In this process, the system obtains the encoder feedback pulse signal of the double-track in real time, and dynamically calculates the position difference and speed difference of the embroidered glass tubes on the two tracks. For example, if the embroidered glass tube on track A is 2% faster than that on track B, the system triggers the camera of track B in advance, or corrects the ROI position of the detection area of track B, to ensure that the embroidered glass tubes on both sides enter the camera field of view at the same time, avoiding missing the crack gap due to incomplete entry of the embroidered glass tube into the field of view.

[0032] The image processing step comprises: the camera transmits the embroidered glass tube images collected in real time to the vision system, the vision system adopts a multi-thread / multi-process parallel processing framework with pipeline splitting of image acquisition, preprocessing, detection and final result output in sequence, allocates independent CPU multi-core and GPU multi-stream processor computing resources to the two tracks respectively, so that the image data processing of the two tracks does not interfere with each other, avoiding blocking of single-track data processing. When the embroidered glass tube on a certain track has a complex crack gap (which requires a longer time for image analysis and defect judgment), the priority scheduling algorithm introduced by the system temporarily increases the computing resource proportion of the track, ensuring that the detection task of the complex crack gap is not delayed, and accurately identifying the crack gap defects of the embroidered glass tube.

[0033] The result processing and sorting, which is specifically completed by the visual system after detecting the broken gap of the embroidery glass tube, uses a self-defined script to process and combine the detection results in real time, plans the corresponding blowing port according to whether there is a broken gap, the size of the gap, the position, etc. When the embroidery glass tube with detection result information reaches the planned blowing position along the double-track conveying system, the system triggers the corresponding blowing port action through the comprehensive blowing pulse position control, blows the defective embroidery glass tube with a broken gap off the track, and conveys the qualified embroidery glass tube to the next process, completing the sorting operation.

[0034] In the embodiment, the detection speed of the double-track visual detection system can reach more than 2500 per minute, which is about 1.8 times higher than the detection speed of 1000-1500 per minute of the traditional single-track, fully meeting the customer's demand for efficient detection of embroidery glass tubes. At the same time, through high-precision hardware configuration and optimized software algorithm, the detection accuracy can reach ±0.02mm, which can accurately detect the broken gap and other defects of the embroidery glass tube. The operation interface is simple and easy to understand, and ordinary workers can operate it. It can also automatically generate production reports, which helps to analyze the proportion of unqualified items and provides a basis for improving the production process of embroidery glass tubes.

[0035] In some embodiments of the application, the way to compensate the detection trigger timing is to adjust the camera exposure time, trigger delay or correct the ROI position of the detection area. The double-track synchronous cooperative control is more operable. Different compensation methods can be flexibly selected according to the actual track deviation, further improving the accuracy of the double-track position alignment and reducing the detection error caused by timing misalignment.

[0036] Through independent computing resources including CPU multi-core and GPU multi-stream processor; the independent computing resources are allocated to the double tracks, including CPU multi-core and GPU multi-stream processor computing resources. The appropriate hardware resources can be accurately selected to ensure the efficiency of parallel processing and ensure real-time data processing without delay.

[0037] The visual detection results are processed and combined in real time by using a self-defined script, and the blowing port is planned. When the corresponding material reaches the planned blowing position, the blowing action is triggered through the comprehensive blowing pulse position control to complete the sorting. The flexibility of the sorting logic and the accuracy of the blowing action are further improved, which can better adapt to the complex sorting needs of different products and improve the accuracy and reliability of the sorting.

[0038] The resolution of the upper and lower detection cameras is not less than 1.3 million pixels, the frame rate is not less than 200 fps, and the lens focal length is 12-50 mm. The response time of the optical fiber sensor is not more than 70 μs, and the detection accuracy is ±0.01 mm. The air pressure adjustment range of the blowing action is 0.2-0.6 MPa, and the blowing time is controlled within 8-12 ms.

[0039] As an example, the following takes the detection of defects in the pins of precision electronic components as an example to further illustrate the content related to the claims of the present application. In this embodiment, the pins of the precision electronic components to be detected are very small, and the clarity, response speed and sorting accuracy of the detection are extremely high.

[0040] By selecting 1.3 million pixels for the upper and lower detection cameras, the frame rate reaches 200 fps, and the lens focal length is selected as 25 mm. Such a configuration allows the subtle defects of the pins of precision electronic components, such as pin bending and deformation, to be clearly captured. The high resolution of 1.3 million pixels ensures the presentation of image details, the frame rate of 35 fps ensures that no detection moment is missed during high-speed conveying on the double track, and the lens focal length of 25 mm precisely focuses on the pin area, avoiding image blur that affects the detection result. The response time of the optical fiber sensor used is 70 μs, and the detection position error is ±0.1 mm. When the precision electronic components are conveyed quickly on the double track, the fast response time of 70 μs can accurately determine the origin of the components, and the detection position error of ±0.1 mm ensures the accuracy of the component position judgment, providing a precise position basis for the subsequent camera shooting trigger, avoiding missed detection or false detection due to position judgment errors. The air pressure adjustment of the blowing action is 0.4 MPa, and the blowing time is controlled within 8 ms. The air pressure of 0.4 MPa can ensure that the precision electronic components with pin defects are accurately blown off the track, and will not cause damage to the components due to excessive air pressure. The blowing time of 30 ms can accurately match the moving speed of the components on the track, ensuring effective sorting when the components pass through the blowing position, and improving the success rate of sorting.

[0041] Through the above configuration, the double-track motion control method can not only accurately detect various pin defects in the detection of defects in the pins of precision electronic components, but also achieve efficient and stable sorting, fully meeting the high requirements of precision electronic component production and detection.

[0042] It should be noted that the above-mentioned custom script is a program written based on a visual programming interface or a specific programming language (such as Python) for defining the logical relationship between the visual detection result and the sorting action. Users can set detection parameter thresholds (such as defect size, number threshold), result judgment logic (such as judging as defective product when multiple defect conditions are met at the same time) and corresponding air outlet selection rules (such as different defect types corresponding to different air outlets) in the script according to the detection needs of different products. After the script is written, it can be imported into the system and compiled and run, and the system processes and combines the visual detection results according to the script logic to generate sorting control instructions. At the same time, the system provides script debugging function, and users can verify the correctness of the script logic through simulation detection data to ensure that the custom script can accurately realize complex sorting logic and meet the patent law requirements for the clarity of technical solutions. The multi-thread / multi-process parallel processing framework adopts a master-slave architecture, the main process is responsible for system resource allocation, task scheduling and communication and coordination between sub-processes / threads; the slave processes / threads are divided into image acquisition threads, preprocessing threads, detection threads and result output threads. Each track corresponds to a group of slave processes / threads, the image acquisition thread acquires image data transmitted by the camera in real time and stores it in the shared memory area; the preprocessing thread reads the image data from the shared memory, performs denoising, enhancement, normalization and other preprocessing operations; the detection thread detects defects and extracts features from the preprocessed image; the result output thread transmits the detection result to the main process and feeds back to the sorting control module. The threads / processes are synchronized and communicated through semaphores and message queues to avoid data competition and resource conflict, ensuring the orderliness and efficiency of parallel processing.

[0043] Although the preferred embodiments of the present application have been described, those skilled in the art who have the basic inventive concept can make further changes and modifications to the embodiments. Therefore, the appended claims are intended to include the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present application.

[0044] Finally, it should be noted that in this document, relational terms such as first and second and the like can only be used to distinguish one entity or action from another entity or action, without necessarily requiring or implying that there is any such actual relationship or order between these entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof are intended to cover non-exclusive inclusions, so that a process, method, article, or terminal device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article, or terminal device. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of additional identical elements in the process, method, article, or terminal device including the element.

[0045] The above describes in detail the dual-rail motion control method provided by the present application. The principles and implementation manners of the present application are described by using specific examples. The above description of the examples is only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will be changed. In summary, the content of the present description should not be understood as a limitation of the present application.

Claims

1. A dual-track motion control method, characterized in that, Includes the following steps: A dual-track conveying system is adopted, with the two tracks spaced 9mm apart. The track surfaces are covered with a hard oxide layer. The glass tray is made of ultra-white high-transmittance glass with a light transmittance of >92%, and the flatness error of the tray surface is controlled within 0.04mm. The drive system uses a DD motor to ensure that the rotation speed of the glass tray is stable at 30-60rpm. Upper and lower inspection cameras are configured to focus on the two ends of the glass tube to detect defects. Point lights are used to illuminate both ends, and the color temperature of the light source is adjusted to 3000K-6500K according to the color of the embroidery texture. Start the equipment and turn on the vibratory feeder when the DD motor reaches a constant speed; The product origin is determined by a fiber optic sensor. When the product passes the set camera position, the camera is triggered to take a picture by combining the high-speed position comparison result. At the same time, the encoder feedback signal of the two tracks is acquired in real time, and the position difference and speed difference of the two tracks are dynamically calculated to compensate for the detection triggering timing. The camera transmits the captured images to the vision system in real time. The vision system adopts a multi-threaded / multi-process parallel processing framework that separates the pipeline from image acquisition to preprocessing, detection, and final output. It allocates independent computing resources to the two tracks and introduces a priority scheduling algorithm. When a track has a complex defect, its resource allocation is temporarily increased to handle complex vision tasks. The visual inspection results are processed and combined in real time using scripts to plan the air blowing ports. When the corresponding material arrives at the planned air blowing position, the air blowing action is triggered to complete the sorting. The feeding process can be selected from dual-track feeding, single-track feeding or dual feeding methods.

2. The dual-track motion control method according to claim 1, characterized in that, The method for compensating for the detection trigger timing is to adjust the camera exposure time, trigger delay, or correct the ROI position of the detection area.

3. The dual-track motion control method according to claim 1, characterized in that, The aforementioned allocation of independent computing resources for dual tracks includes allocating CPU multi-core and GPU multi-stream processor computing resources for dual tracks.

4. The dual-track motion control method according to claim 1, characterized in that, The process of real-time scripting and combining visual inspection results to plan air inlets, and triggering an air blowing action when the corresponding material reaches the planned air blowing position, completes the sorting; includes: A custom script is used to process and combine the visual inspection results in real time, and the air blowing port is planned. When the corresponding material arrives at the planned air blowing position, the air blowing action is triggered by the comprehensive air blowing pulse position control to complete the sorting.

5. The dual-track motion control method according to claim 1, characterized in that, The resolution of both the upper and lower detection cameras is no less than 1.3 million pixels, the frame rate is no less than 200fps, and the lens focal length is 12-50mm.

6. The dual-track motion control method according to claim 1, characterized in that, The response time of the fiber optic sensor is no more than 70 μs, and the detection position error is ±0.1 mm.

7. The dual-track motion control method according to claim 1, characterized in that, The air pressure adjustment range for the blowing action is 0.2-0.6MPa, and the blowing time is controlled between 4-12ms.

Images