AI platform acquisition system

Abstract

The invention discloses an AI platform acquisition system, and relates to the technical field of image acquisition, and aims to solve the problem that multi-direction automatic acquisition of tools cannot be realized in the prior art, the AI platform acquisition system comprises a rotation control module which calculates the accelerated feed amount of a motor according to operation parameters, analyzes the relationship between the accelerated feed amount and the minimum acquisition distance between adjacent acquired images, and outputs the acquired images; comparing the actual acquisition distance between the adjacent images with the theoretical feed amount of the motor, and determining a feed error; the rotation adjusting module is used for carrying out compensation adjustment according to the obtained feeding error; the acquisition module is used for acquiring image information of the tools on the turntable through a plurality of cameras; and the image optimization module is used for carrying out image splicing on the adjacent acquired images and optimizing the spliced images. According to the AI platform acquisition system, automatic acquisition and processing of images can be realized, and the intelligent identification level of tools and instruments is improved.

Description

Technical Field

[0001] The present invention relates to the field of image acquisition technology, and in particular to an AI platform acquisition system. Background Art

[0002] Existing tool image acquisition is generally divided into manual and machine. Manual acquisition usually uses mobile phones, cameras and other recording devices. The pixels of mobile phones cannot meet the requirements and the memory and storage space are small. Camera operation is difficult, time-consuming and labor-intensive. Once the number of targets to be acquired increases, manual acquisition cannot meet the work requirements. Machine acquisition generally uses monitoring equipment, which has low pixels and the shooting angle needs to be manually adjusted, which does not completely save manpower.

[0003] The Chinese patent with publication number CN110567973B, which was published on July 5, 2022, relates to a piston monitoring platform and method based on image acquisition, and discloses an image acquisition platform, the upper surface of which is provided with a rotating mechanism, which can clamp the piston and drive the piston to rotate, and the upper part is provided with multiple image acquisition mechanisms, which are arranged around the rotating mechanism. The image acquisition mechanism is composed of several cameras, which can be moved vertically to adjust the focal length, and the distance between the camera and the device can also be adjusted to adapt to different device specifications. However, this patent cannot realize multi-directional automatic acquisition of tools. Summary of the invention

[0004] The present invention solves the problem that the prior art cannot realize multi-directional automatic acquisition of tools and instruments, and proposes an AI platform acquisition system that can realize automatic acquisition and processing of images and improve the intelligent recognition level of tools and instruments.

[0005] In order to achieve the above object, the present invention adopts the following technical solution: an AI platform acquisition system, comprising: The rotation control module calculates the acceleration feed amount of the motor according to the operating parameters, analyzes the relationship between the acceleration feed amount and the minimum acquisition spacing between adjacent acquisition images, and compares the actual acquisition spacing between adjacent images with the theoretical feed amount of the motor to determine the feed error; A rotation adjustment module performs compensation adjustment according to the obtained feed error; The acquisition module collects image information of the tools on the turntable through several cameras; The image optimization module performs image stitching on adjacent acquired images and optimizes the stitched images.

[0006] The technical solution mainly includes a rotation control module, a rotation adjustment module, an acquisition module and an image optimization module. The rotation control module can control the motor that drives the turntable, and obtain the accelerated feed amount and the theoretical feed amount of the motor by analyzing and calculating the operating parameters of the motor, and finally obtain the feed error; then the motor feed amount is compensated and adjusted in the rotation adjustment module; the acquisition module can acquire image information of the rotating tool; the image optimization module can realize image stitching and optimization of adjacent images.

[0007] The present invention is further configured as follows: the rotation control module comprises: A motor parameter acquisition unit is used to acquire the motor's starting speed and the number of pulses during the starting process; The rotation parameter analysis unit analyzes the motor rotation acceleration within a unit time period according to the motor rotation speed corresponding to a single pulse; The rotation comparison unit determines whether the accelerated feed amount is less than the minimum acquisition spacing between adjacent acquired images and determines the feed error; Error analysis unit, which compensates the motor running time based on feed error analysis.

[0008] In the present technical solution, the motor parameter acquisition unit is connected to the rotation parameter analysis unit, the rotation parameter analysis unit is connected to the rotation comparison unit, and the rotation comparison unit is connected to the error analysis unit. After the motor parameter acquisition unit acquires the corresponding motor parameters, the rotation parameter analysis unit immediately starts parameter analysis to obtain the motor rotation acceleration within a unit time period; then, the rotation comparison unit continues to analyze to obtain the motor rotation amount and the motor acceleration feed amount, determines whether the acceleration feed amount is less than the minimum acquisition interval, and analyzes to obtain the corresponding feed error; finally, the error analysis unit specifically analyzes the compensation amount of the motor running time, so as to perform adjustment and compensation in the rotation adjustment module later.

[0009] The present invention is further configured as follows: the rotation adjustment module comprises: an error acquisition unit, receiving a feed error from a rotation control module and compensating for a motor running time; The adjustment and compensation unit is connected to the error acquisition unit to perform compensation and adjustment on the motor.

[0010] In this technical solution, the turntable motor is compensated and adjusted according to the feed error result of the rotation control module and the compensation motor running time obtained by analysis, thereby ensuring the matching of the turntable rotation and image acquisition.

[0011] The present invention is further configured as follows: the acquisition module includes a plurality of cameras and an image preprocessing unit connected to the plurality of cameras, and the image preprocessing unit can classify and divide the acquired images according to tools to obtain image information of the same tool in different directions.

[0012] In this technical solution, since the tools on the turntable are rotating, the acquisition module can be used to obtain image information about the tools in different directions. At least two tool images will be captured for the same tool to obtain more complete tool information.

[0013] The present invention is further configured as follows: the image optimization module comprises: An optimization unit performs local enhancement processing on the stitched images; For the overlap unit, when the acceleration feed amount is less than the minimum acquisition spacing between adjacent acquired images, the overlap stitching strategy is used to generate a stitched image; AI unit, when the acceleration feed amount is greater than or equal to the minimum acquisition spacing between adjacent acquired images, uses AI predictive stitching technology to perform image stitching.

[0014] In this technical solution, the overlap unit and the AI ​​unit are selectively selected, and both are connected to the optimization unit. After the stitching is completed, the image is optimized by the optimization unit to ensure subsequent operability.

[0015] The present invention is further configured as follows: the rotation control module also includes a feedback unit and a database connected to the feedback unit, the feedback unit is capable of receiving the image optimization results of the image optimization module and storing the image optimization results in the database, and the database also includes process data in each unit of the rotation control module.

[0016] In this technical solution, the feedback unit can first screen the image optimization results so that the optimization result data entering the database are all reference data.

[0017] The present invention is further configured as follows: the overlap unit comprises: Cutting sub-units, extracting overlapping areas of adjacent images, and determining several feature points of the overlapping areas; The grayscale correction subunit corrects the grayscale values ​​of the feature points in the overlapping area according to the distribution of the feature points in the two images.

[0018] In this technical solution, overlapping and splicing are performed by cutting out overlapping areas to ensure complete recording of tool images.

[0019] The present invention is further configured as follows: an AI prediction model is provided in the AI ​​unit, and the AI ​​prediction model is trained based on the AI ​​prediction splicing technology. The AI ​​prediction model can predict partial features of tools in the blank acquisition area between two acquired images, thereby forming a spliced ​​image.

[0020] In this technical solution, AI stitching is aimed at the case where the accelerated feed amount is greater than or equal to the minimum acquisition spacing between adjacent acquired images. Since there will be a blank acquisition area between two adjacent acquired images in this case, AI technology is used for stitching.

[0021] The present invention is further configured as follows: the database is also connected to a rotation parameter analysis unit, and the error analysis unit is connected to an error acquisition unit of a rotation adjustment module.

[0022] In this technical solution, the data in the database can provide data support for the rotation parameter analysis process.

[0023] The present invention is further configured as follows: the rotation adjustment module and the rotation control module are both arranged in a control cabinet, a turntable is arranged above the control cabinet, and a collection module is arranged above the turntable.

[0024] In the technical solution, the rotation adjustment module and the rotation control module can be integrated or separately arranged, and both are arranged in the control cabinet.

[0025] The present invention can bring the following beneficial effects: The present invention relates to an AI platform acquisition system that can accurately control a turntable in a straight line, ensure multi-angle shooting of tool images, and optimize spliced ​​images; it can realize automatic acquisition and processing of images, and improve the intelligent recognition level of tools. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a schematic diagram of the principle of an AI platform acquisition system of the present application.

[0027] Figure 2 It is a structural diagram of an AI platform acquisition system of the present application.

[0028] Reference numerals 1. Acquisition module 2. Control cabinet 3. Turntable. DETAILED DESCRIPTION

[0029] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0030] Example 1 This embodiment proposes an AI platform acquisition system, referring to Figure 1 and Figure 2 It mainly includes a rotation control module, a rotation adjustment module, a collection module 1 and an image optimization module. The rotation control module and the rotation adjustment module are both arranged in a control cabinet 2. A turntable 3 is arranged above the control cabinet 2, and the collection module 1 is arranged above the turntable 3.

[0031] For the rotation control module, it can mainly calculate the acceleration feed amount of the motor according to the operating parameters, and then analyze the relationship between the acceleration feed amount and the minimum acquisition spacing between adjacent acquired images, and compare the actual acquisition spacing between adjacent images with the theoretical feed amount of the motor to determine the feed error.

[0032] refer to Figure 1 The rotation control module mainly includes the following units.

[0033] The motor parameter acquisition unit can mainly acquire the starting speed of the motor and the number of pulses during the starting process; in this embodiment, the number of pulses during the starting process is mainly the number of pulses in the set unit time from the starting to the target state of the motor; the target state is the state of reaching the target speed.

[0034] The rotation parameter analysis unit can mainly analyze the motor rotation acceleration within a unit time period according to the motor rotation speed corresponding to a single pulse; specifically, according to the motor rotation speed corresponding to a single pulse and the number of pulses, the motor rotation acceleration within a unit time period is finally obtained.

[0035] The rotation comparison unit can determine whether the accelerated feed amount is less than the minimum acquisition spacing between adjacent acquired images and determine the feed error; more specifically, it can calculate the motor rotation amount and the motor accelerated feed amount based on the result of the rotation parameter analysis unit, and then determine whether the accelerated feed amount of the motor from starting quick freezing to reaching the target state is less than the preset minimum acquisition spacing between adjacent acquired images, and calculate the feed error at the same time.

[0036] The error analysis unit can compensate the motor running time according to the feed error analysis. Specifically, the motor running time is compensated according to the feed error to obtain the compensation time of the motor in the adjacent image acquisition interval.

[0037] In the above technical solution, the motor parameter acquisition unit is connected to the rotation parameter analysis unit, the rotation parameter analysis unit is connected to the rotation comparison unit, and the rotation comparison unit is connected to the error analysis unit. After the motor parameter acquisition unit obtains the corresponding motor parameters, the rotation parameter analysis unit immediately starts parameter analysis to obtain the motor rotation acceleration within a unit time period; then the rotation comparison unit continues to analyze to obtain the motor rotation amount and the motor acceleration feed amount, determines whether the acceleration feed amount is less than the minimum acquisition interval, and analyzes to obtain the corresponding feed error; finally, the error analysis unit specifically analyzes the compensation amount of the motor running time, so as to perform adjustment and compensation in the rotation adjustment module later.

[0038] refer to Figure 1 As for the acquisition module, it mainly includes a number of cameras, which are used to acquire image information of the tools on the turntable.

[0039] More specifically, the acquisition module mainly includes several cameras and an image preprocessing unit. Each camera is connected to the image preprocessing unit, and the image preprocessing unit can divide the image. The specific classification is based on the type of tool obtained by the acquisition. After the division, the image information of the same tool will be stored together for subsequent processing.

[0040] In the above technical solution, since the tools on the turntable are rotating, the acquisition module can be used to obtain image information about the tools in different directions. At least two tool images will be captured for the same tool to obtain more complete tool information.

[0041] Continue to refer Figure 1 For the rotation adjustment module, it can mainly perform compensation adjustment operations based on the obtained feed error.

[0042] The rotation adjustment module mainly includes an error acquisition unit and an adjustment compensation unit. The error acquisition unit and the adjustment compensation unit are connected to each other. The error acquisition unit can directly receive the feed error in the rotation control module and compensate for the motor running time; the adjustment compensation unit can perform compensation adjustment based on the data obtained by the error acquisition unit.

[0043] In the above technical solution, the turntable motor is compensated and adjusted according to the feed error result of the rotation control module and the compensation motor running time obtained by analysis, thereby ensuring the matching of the turntable rotation and image acquisition.

[0044] The image optimization module mainly performs the functions of stitching adjacent acquired images and optimizing the stitched images.

[0045] More specifically, the image optimization module mainly includes an optimization unit, an overlap unit and an AI unit.

[0046] The overlap unit is mainly used for the situation where the acceleration feed amount is less than the minimum acquisition spacing between adjacent acquired images. Specifically, an overlap stitching strategy is used to generate a stitched image.

[0047] The AI ​​unit is mainly used for situations where the acceleration feed amount is greater than or equal to the minimum acquisition spacing between adjacent acquired images. Specifically, it uses AI predictive stitching technology to perform image stitching.

[0048] The function performed by the optimization unit is to perform local enhancement processing on the stitched images.

[0049] In the above technical solution, the overlap unit and the AI ​​unit are selected one by one, and both are connected to the optimization unit. After the stitching is completed, the image is optimized by the optimization unit to ensure subsequent operability.

[0050] Continue to refer Figure 1 The rotation control module also includes a feedback unit and a database. The database and the feedback unit are interconnected. The feedback unit can receive the image optimization results of the image optimization module and store the image optimization results in the database. The database also includes process data in each unit of the rotation control module.

[0051] In the above technical solution, the feedback unit can first screen the image optimization results so that the optimization result data entering the database are all reference data.

[0052] For the above-mentioned overlap unit, more specifically, it mainly includes a cut-off subunit and a grayscale correction subunit. The cut-off subunit can mainly extract the overlapping area of ​​adjacent images and determine a number of feature points in the overlapping area; the grayscale correction subunit can correct the grayscale values ​​of the feature points in the overlapping area according to the distribution of the feature points in the two images.

[0053] In the above technical solution, overlapping and splicing are performed by cutting out the overlapping areas to ensure the complete recording of the tool image.

[0054] The overlap stitching strategy implemented by the overlap unit is to first extract the overlap area of ​​adjacent images and determine several feature points in the overlap area; then the grayscale values ​​of the feature points in the overlap area are corrected according to the distribution of the feature points in the two images.

[0055] The above-mentioned AI unit has an AI prediction model disposed therein. The AI ​​prediction model is trained based on the AI ​​prediction stitching technology. The AI ​​prediction model can predict the partial features of the tool in the blank acquisition area between two acquired images, and then form a stitched image.

[0056] In the above technical solution, AI stitching is aimed at the case where the accelerated feed amount is greater than or equal to the minimum acquisition spacing between adjacent acquired images. Since there will be a blank acquisition area between two adjacent acquired images in this case, AI technology is used for stitching.

[0057] The rotation adjustment module and the rotation control module are both arranged in a control cabinet. A turntable is arranged above the control cabinet, and a collection module is arranged above the turntable.

[0058] In the above technical solution, the rotation adjustment module and the rotation control module can be integrated or separately arranged, and both are arranged in a control cabinet.

[0059] An AI platform acquisition system of the present technical solution mainly includes a rotation control module, a rotation adjustment module, an acquisition module and an image optimization module. The rotation control module can control the motor that drives the turntable, and obtain the accelerated feed amount of the motor and the theoretical feed amount of the motor by analyzing and calculating the operating parameters of the motor, and finally obtain the feed error; then the motor feed amount is compensated and adjusted in the rotation adjustment module; the acquisition module can acquire image information of rotating tools; the image optimization module can realize image stitching and optimization of adjacent images; the AI ​​platform acquisition system of the present application can realize automatic acquisition and processing of images, and can improve the intelligent recognition level of tools.

[0060] Example 2 This embodiment also proposes an AI platform acquisition system, referring to Figure 1 and Figure 2 It mainly includes a rotation control module, a rotation adjustment module, a collection module 1 and an image optimization module. The rotation control module and the rotation adjustment module are both arranged in a control cabinet 2. A turntable 3 is arranged above the control cabinet 2, and the collection module 1 is arranged above the turntable 3.

[0061] For the rotation control module, it can mainly calculate the acceleration feed amount of the motor according to the operating parameters, and then analyze the relationship between the acceleration feed amount and the minimum acquisition spacing between adjacent acquired images, and compare the actual acquisition spacing between adjacent images with the theoretical feed amount of the motor to determine the feed error.

[0062] refer to Figure 1 The rotation control module mainly includes the following units.

[0063] The motor parameter acquisition unit can mainly acquire the starting speed of the motor and the number of pulses during the starting process; in this embodiment, the number of pulses during the starting process is mainly the number of pulses in the set unit time from the starting to the target state of the motor; the target state is the state of reaching the target speed.

[0064] The rotation parameter analysis unit can mainly analyze the motor rotation acceleration within a unit time period according to the motor rotation speed corresponding to a single pulse; specifically, according to the motor rotation speed corresponding to a single pulse and the number of pulses, the motor rotation acceleration within a unit time period is finally obtained.

[0065] The rotation comparison unit can determine whether the accelerated feed amount is less than the minimum acquisition spacing between adjacent acquired images and determine the feed error; more specifically, it can calculate the motor rotation amount and the motor accelerated feed amount based on the result of the rotation parameter analysis unit, and then determine whether the accelerated feed amount of the motor from starting quick freezing to reaching the target state is less than the preset minimum acquisition spacing between adjacent acquired images, and calculate the feed error at the same time.

[0066] The error analysis unit can compensate the motor running time according to the feed error analysis. Specifically, the motor running time is compensated according to the feed error to obtain the compensation time of the motor in the adjacent image acquisition interval.

[0067] In the above technical solution, the motor parameter acquisition unit is connected to the rotation parameter analysis unit, the rotation parameter analysis unit is connected to the rotation comparison unit, and the rotation comparison unit is connected to the error analysis unit. After the motor parameter acquisition unit obtains the corresponding motor parameters, the rotation parameter analysis unit immediately starts parameter analysis to obtain the motor rotation acceleration within a unit time period; then the rotation comparison unit continues to analyze to obtain the motor rotation amount and the motor acceleration feed amount, determines whether the acceleration feed amount is less than the minimum acquisition interval, and analyzes to obtain the corresponding feed error; finally, the error analysis unit specifically analyzes the compensation amount of the motor running time, so as to perform adjustment and compensation in the rotation adjustment module later.

[0068] refer to Figure 1 As for the acquisition module, it mainly includes a number of cameras, which are used to acquire image information of the tools on the turntable.

[0069] More specifically, the acquisition module mainly includes several cameras and an image preprocessing unit. Each camera is connected to the image preprocessing unit, and the image preprocessing unit can divide the image. The specific classification is based on the type of tool obtained by the acquisition. After the division, the image information of the same tool will be stored together for subsequent processing.

[0070] In the above technical solution, since the tools on the turntable are rotating, the acquisition module can be used to obtain image information about the tools in different directions. At least two tool images will be captured for the same tool to obtain more complete tool information.

[0071] Continue to refer Figure 1 For the rotation adjustment module, it can mainly perform compensation adjustment operations based on the obtained feed error.

[0072] The rotation adjustment module mainly includes an error acquisition unit and an adjustment compensation unit. The error acquisition unit and the adjustment compensation unit are connected to each other. The error acquisition unit can directly receive the feed error in the rotation control module and compensate for the motor running time; the adjustment compensation unit can perform compensation adjustment based on the data obtained by the error acquisition unit.

[0073] In the above technical solution, the turntable motor is compensated and adjusted according to the feed error result of the rotation control module and the compensation motor running time obtained by analysis, thereby ensuring the matching of the turntable rotation and image acquisition.

[0074] The image optimization module mainly performs the functions of stitching adjacent acquired images and optimizing the stitched images.

[0075] More specifically, the image optimization module mainly includes an optimization unit, an overlap unit and an AI unit.

[0076] The overlap unit is mainly used for the situation where the acceleration feed amount is less than the minimum acquisition spacing between adjacent acquired images. Specifically, an overlap stitching strategy is used to generate a stitched image.

[0077] The AI ​​unit is mainly used for situations where the acceleration feed amount is greater than or equal to the minimum acquisition spacing between adjacent acquired images. Specifically, it uses AI predictive stitching technology to perform image stitching.

[0078] The function performed by the optimization unit is to perform local enhancement processing on the stitched images.

[0079] In the above technical solution, the overlap unit and the AI ​​unit are selected one by one, and both are connected to the optimization unit. After the stitching is completed, the image is optimized by the optimization unit to ensure subsequent operability.

[0080] Continue to refer Figure 1 The rotation control module also includes a feedback unit and a database. The database and the feedback unit are interconnected. The feedback unit can receive the image optimization results of the image optimization module and store the image optimization results in the database. The database also includes process data in each unit of the rotation control module.

[0081] In the above technical solution, the feedback unit can first screen the image optimization results so that the optimization result data entering the database are all reference data.

[0082] For the above-mentioned overlap unit, more specifically, it mainly includes a cut-off subunit and a grayscale correction subunit. The cut-off subunit can mainly extract the overlapping area of ​​adjacent images and determine a number of feature points in the overlapping area; the grayscale correction subunit can correct the grayscale values ​​of the feature points in the overlapping area according to the distribution of the feature points in the two images.

[0083] In the above technical solution, overlapping and splicing are performed by cutting out the overlapping areas to ensure the complete recording of the tool image.

[0084] The overlap stitching strategy implemented by the overlap unit is to first extract the overlap area of ​​adjacent images and determine several feature points in the overlap area; then the grayscale values ​​of the feature points in the overlap area are corrected according to the distribution of the feature points in the two images.

[0085] The above-mentioned AI unit has an AI prediction model disposed therein. The AI ​​prediction model is trained based on the AI ​​prediction stitching technology. The AI ​​prediction model can predict the partial features of the tool in the blank acquisition area between two acquired images, and then form a stitched image.

[0086] In the above technical solution, AI stitching is aimed at the case where the accelerated feed amount is greater than or equal to the minimum acquisition spacing between adjacent acquired images. Since there will be a blank acquisition area between two adjacent acquired images in this case, AI technology is used for stitching.

[0087] The rotation adjustment module and the rotation control module are both arranged in a control cabinet. A turntable is arranged above the control cabinet, and a collection module is arranged above the turntable.

[0088] In the above technical solution, the rotation adjustment module and the rotation control module can be integrated or separately arranged, and both are arranged in a control cabinet.

[0089] Different from Example 1, the database is also connected to the rotation parameter analysis unit, and the error analysis unit is connected to the error acquisition unit of the rotation adjustment module. In the above technical solution, the data in the database can provide data support for the rotation parameter analysis process, and can further ensure the analysis accuracy of the rotation parameter analysis unit.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention is described in detail with reference to the embodiments, it should be understood by those skilled in the art that any modification or equivalent replacement of the technical solutions of the present invention does not depart from the spirit and scope of the technical solutions of the present invention and should be included in the scope of the claims of the present invention.

Claims

1. An AI platform acquisition system, characterized in that: include: The rotation control module calculates the acceleration feed amount of the motor according to the operating parameters, analyzes the relationship between the acceleration feed amount and the minimum acquisition spacing between adjacent acquisition images, and compares the actual acquisition spacing between adjacent images with the theoretical feed amount of the motor to determine the feed error; A rotation adjustment module performs compensation adjustment according to the obtained feed error; The acquisition module collects image information of the tools on the turntable through several cameras; The image optimization module performs image stitching on adjacent acquired images and optimizes the stitched images.

2. The AI ​​platform acquisition system according to claim 1, characterized in that: The rotation control module comprises: A motor parameter acquisition unit is used to acquire the motor's starting speed and the number of pulses during the starting process; The rotation parameter analysis unit analyzes the motor rotation acceleration within a unit time period according to the motor rotation speed corresponding to a single pulse; The rotation comparison unit determines whether the accelerated feed amount is less than the minimum acquisition spacing between adjacent acquired images and determines the feed error; Error analysis unit, which compensates the motor running time based on feed error analysis.

3. An AI platform acquisition system according to claim 1 or 2, characterized in that: The rotation adjustment module comprises: an error acquisition unit, receiving a feed error from a rotation control module and compensating for a motor running time; The adjustment and compensation unit is connected to the error acquisition unit to perform compensation and adjustment on the motor.

4. The AI ​​platform acquisition system according to claim 3, characterized in that: The acquisition module includes a plurality of cameras and an image preprocessing unit connected to the plurality of cameras. The image preprocessing unit can classify and divide the acquired images according to the tools to obtain image information of the same tool in different directions.

5. An AI platform acquisition system according to claim 1, 2 or 4, characterized in that: The image optimization module comprises: An optimization unit performs local enhancement processing on the stitched image; For the overlap unit, when the acceleration feed amount is less than the minimum acquisition spacing between adjacent acquired images, the overlap stitching strategy is used to generate a stitched image; AI unit, when the acceleration feed amount is greater than or equal to the minimum acquisition spacing between adjacent acquired images, uses AI predictive stitching technology to perform image stitching.

6. The AI ​​platform acquisition system according to claim 2, characterized in that: The rotation control module also includes a feedback unit and a database connected to the feedback unit. The feedback unit can receive the image optimization result of the image optimization module and store the image optimization result in the database. The database also includes process data in each unit of the rotation control module.

7. The AI ​​platform acquisition system according to claim 5, characterized in that: The overlap unit comprises: Cutting sub-units, extracting overlapping areas of adjacent images, and determining several feature points of the overlapping areas; The grayscale correction subunit corrects the grayscale values ​​of the feature points in the overlapping area according to the distribution of the feature points in the two images.

8. The AI ​​platform acquisition system according to claim 5, characterized in that: The AI ​​unit is provided with an AI prediction model, which is trained based on the AI ​​prediction splicing technology. The AI ​​prediction model can predict the partial features of the tool in the blank acquisition area between two acquired images, and then form a spliced ​​image.

9. The AI ​​platform acquisition system according to claim 6, characterized in that: The database is also connected to a rotation parameter analysis unit, and the error analysis unit is connected to an error acquisition unit of a rotation adjustment module.

10. An AI platform acquisition system according to claim 1, 2 or 4, characterized in that: The rotation adjustment module and the rotation control module are both arranged in a control cabinet. A turntable is arranged above the control cabinet, and a collection module is arranged above the turntable.

Images