A robot for processing the inner wall of a steel cylinder
By using intelligent chips and deep learning models in the cylinder inner wall processing robot, targeted detection solutions are generated and efficient texture feature data are extracted, which solves the problems of low detection efficiency, inaccurate image acquisition and low degree of automation of texture feature extraction in the prior art. , realizes efficient and accurate detection of inner wall of steel cylinder and corrosion treatment.
Patent Information
- Application Number
- CN202411088594.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-09
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-09
AI Technical Summary
The lack of targeted detection during the processing of the inner wall of the existing cylinders, resulting in inefficient efficiency, inaccurate image acquisition direction, increased detection difficulty, low automation of texture feature data extraction, and difficult to accurately identify inner wall defects and remove rust.
Design a cylinder inner wall processing robot, including two robotic arms and smart chips, obtain cylinder information through positioning sensors and label reading devices, generate targeted detection solutions, and use deep learning models and image acquisition indicators (such as longitudinal shooting height and horizontal directions) Rotate angle) perform image acquisition and extract efficient texture feature data by segmenting image blocks.
It improves detection efficiency and accuracy, saves resources and costs, and can fully capture specific parts of the inner wall of the cylinder, quickly identify inner wall defects and remove rust in time, avoiding potential safety risks caused by rust.
Smart Images

Figure CN118596124B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of robots, and in particular to a steel cylinder inner wall processing robot. Background Art
[0002] With the continuous development of industrial automation technology, all walks of life are seeking ways to improve production efficiency, reduce manual participation and ensure product quality. In recent years, cylinder inner wall processing robots have been widely used in the industrial field. Traditional manual processing methods can no longer meet the efficiency and accuracy requirements of modern production lines. Cylinder inner wall processing robots can independently complete the detection task of cylinder inner wall information without manual intervention, which improves production efficiency and accuracy.
[0003] The existing cylinder inner wall processing process, especially the cylinder inner wall inspection process, often has the following technical problems:
[0004] First, traditional cylinder inspection is extensive but not targeted. All cylinders that may be defective are checked, which increases the time and resources required for cylinder inspection. The lack of targeted inspection leads to reduced efficiency.
[0005] Second, during the image acquisition process of the inner wall of the cylinder, due to the inaccurate shooting direction, the actual condition of the inner wall of the cylinder cannot be fully reflected, and the low-resolution images of key areas may increase the difficulty of maintenance and inspection;
[0006] Third, the traditional method of extracting texture feature data from the inner wall of the steel cylinder has a low degree of automation, making it difficult to extract effective texture feature data locally. The resulting texture feature results are inefficiently represented, causing difficulties in subsequent processing. It is impossible to accurately identify all types of inner wall defects, nor is it possible to remove rust on the inner wall of the steel cylinder in a timely manner. Summary of the invention
[0007] The summary of the invention is used to introduce concepts in a brief form, which will be described in detail in the detailed implementation section below. The summary of the invention is not intended to identify the key features or essential features of the technical solution claimed for protection, nor is it intended to limit the scope of the technical solution claimed for protection.
[0008] The present invention proposes a cylinder inner wall processing robot to solve one or more of the technical problems mentioned in the above background technology part.
[0009] The present invention provides a steel cylinder inner wall processing robot, comprising: two mechanical arms, the two mechanical arms cooperate with each other to place the steel cylinder to be detected on a detection table, the end of the first mechanical arm of the two mechanical arms is provided with an image acquisition device, and the end of the second mechanical arm of the two mechanical arms is provided with a label reading device; a detection table, the detection table is used to place the steel cylinder to be detected, and a plurality of positioning sensors and positioning marks are provided on the detection table. When the steel cylinder to be detected is placed at a target position corresponding to the positioning mark, the plurality of positioning sensors send prompt information indicating that the placement of the steel cylinder to be detected is completed to an intelligent chip; an intelligent chip, the intelligent chip is used to control the second mechanical arm to drive the label reading device to read the steel cylinder information in the electronic tag set on the steel cylinder to be detected after receiving the prompt information, and parse the steel cylinder information to obtain a plurality of attribute information and historical maintenance information of the steel cylinder to be detected, the plurality of attribute information including production date, production batch, and steel cylinder type, and the historical maintenance information includes a plurality of maintenance records, each maintenance record including maintenance time, Maintenance location and maintenance method; the smart chip is also used to determine the service life of the steel cylinder to be inspected according to the production date, and query the multiple inspection items corresponding to the steel cylinder to be inspected in the pre-configured inspection item query table according to the service life, and the multiple inspection items queried constitute a candidate inspection item set; query the multiple inner wall defects corresponding to the steel cylinders in the production batch and the historical number of detections corresponding to each inner wall defect in the defect detection database according to the production batch; select the inner wall defect with a historical number of detections greater than the preset number of detections from the multiple inner wall defects as the inner wall defect to be checked, and add the inspection items corresponding to the inner wall defect to be checked into the candidate inspection item set to obtain a standard inspection item set; the smart chip is also used to generate an inner wall inspection plan based on the standard inspection item set, the inner wall inspection plan includes image acquisition indicators, and the first mechanical arm is controlled according to the inner wall inspection plan to drive the image acquisition device to collect images of the inner wall of the steel cylinder to be inspected to obtain an inner wall image set, and the inner wall image set is used to process the inner wall of the steel cylinder to be inspected.
[0010] Optionally, the image acquisition index includes a vertical shooting height and a horizontal rotation angle, and the vertical shooting height and the horizontal rotation angle are generated by the following steps:
[0011] According to the cylinder type, a cylinder partitioning scheme library that is pre-configured is matched to obtain a cylinder partitioning scheme corresponding to the cylinder to be detected. The cylinder partitioning scheme includes the number of axial partitions and the number of lateral partitions.
[0012] According to the number of axial partitions, the cylinder to be inspected is axially divided to obtain multiple axial partitions, and the axial height interval corresponding to each axial partition is determined;
[0013] According to the number of transverse partitions, each of the multiple axial partitions is transversely divided to obtain multiple transverse partitions corresponding to each axial partition, and the rotation angle interval corresponding to each transverse partition is determined, and the transverse partitions corresponding to the multiple axial partitions respectively form a transverse partition set;
[0014] According to the axial height interval corresponding to each axial partition and the rotation angle interval corresponding to each lateral partition, the image acquisition index corresponding to each lateral partition in the lateral partition set is determined, and the image acquisition index includes the longitudinal shooting height and the lateral rotation angle.
[0015] Optionally, the image acquisition indicator includes an image acquisition resolution, and the image acquisition resolution is generated by the following steps:
[0016] Generate a re-inspection score corresponding to each maintenance record in multiple maintenance records according to the maintenance time and maintenance method. If the re-inspection score is greater than a preset re-inspection score threshold, determine the maintenance location contained in the corresponding maintenance record as the target maintenance location, and match the target maintenance location in the horizontal partition set to obtain a horizontal partition matching the target maintenance location and use it as the key horizontal partition;
[0017] A first image acquisition resolution is configured for a key horizontal partition in the horizontal partition set, and a second image acquisition resolution is configured for the remaining horizontal partitions in the horizontal partition set except the key horizontal partition, wherein the first image acquisition resolution is greater than the second image acquisition resolution.
[0018] Optionally, the inner wall detection scheme is generated by the following steps:
[0019] Match the deep learning model corresponding to each detection item from the pre-configured deep learning model library to obtain a deep learning model set corresponding to the standard detection item set;
[0020] A collection of deep learning models and image acquisition metrics form an inner wall detection solution.
[0021] The present invention has the following beneficial effects:
[0022] 1. The test table is equipped with multiple positioning sensors and positioning marks, which improves the accuracy of placing the steel cylinder to be tested at the target position and enhances the detection efficiency. In addition, the smart chip queries the multiple detection items corresponding to the steel cylinder to be tested in the pre-configured detection item query table according to the service life. The service life can avoid unnecessary detection of steel cylinders that have not reached the detection cycle, ensuring the targeted detection, thereby saving detection resources and costs. According to the production batch, the multiple inner wall defects corresponding to the steel cylinders of the production batch and the historical detection times corresponding to each inner wall defect can be queried in the defect detection database, so as to quickly understand the inner wall defects of the steel cylinders of the production batch, conduct targeted detection, and improve the detection efficiency. According to the standard detection item set, the inner wall detection plan is generated. By following the standard detection items, the scientific nature of the generated inner wall detection plan can be ensured.
[0023] 2. The vertical shooting height and horizontal rotation angle can accurately capture specific parts of the cylinder and can fully capture any area of the cylinder, improving the efficiency of subsequent image analysis; higher image acquisition resolution is configured for key target maintenance locations, so that the collected images can provide more information, thereby obtaining a more scientific and accurate inner wall inspection solution;
[0024] 3. In the process of extracting texture feature data by the intelligent chip, the efficiency of the local processing process is improved by dividing the image into multiple image blocks; by comparing the neighborhood pixels with the central pixels, the changes in local texture features can be captured, and the encoding obtained by comparing the domain pixels with the central pixels is a binary number, which provides an efficient texture feature representation method, which is convenient for the subsequent texture feature recognition of the image. The calculation is simple, efficient and highly automated. The pre-trained network can quickly process the input texture feature data and quickly identify the types of various inner wall defects. Timely rust removal can avoid potential safety risks caused by rust. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.
[0026] Figure 1 It is a structural schematic diagram of a steel cylinder inner wall processing robot of the present invention. DETAILED DESCRIPTION
[0027] The present invention will be described in more detail below with reference to the accompanying drawings. Although certain 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 construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not intended to limit the scope of protection of the present invention.
[0028] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other.
[0029] It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.
[0030] It should be noted that the modifications of "one" and "plurality" mentioned in the present invention are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".
[0031] The names of the messages or information exchanged between multiple devices of the present invention are only for illustrative purposes, and are not used to limit the scope of these messages or information.
[0032] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.
[0033] like Figure 1 As shown, a schematic diagram of the structure of a cylinder inner wall processing robot of the present invention. The cylinder inner wall processing robot includes two mechanical arms 101, a detection platform 102 and an intelligent chip 103. The two mechanical arms 101 cooperate with each other to place the cylinder to be tested on the detection platform. In some embodiments, the mechanical arm is a structure that imitates human arms. The mechanical arm places the cylinder to be tested on the detection platform. The two mechanical arms are composed of multiple joints and connected parts. Each of the two mechanical arms is connected to a mechanical gripper. The two mechanical grippers at the ends of the two mechanical arms synchronously grip the cylinder to prevent the cylinder from sliding or falling during the gripping process. An image acquisition device is provided at the end of the first of the two mechanical arms, and a label reading device is provided at the end of the second of the two mechanical arms. The image acquisition device can be a camera, and the label reading device can be an RFID reader.
[0034] The testing platform 102 is used to place the cylinder to be tested. A plurality of positioning sensors and positioning marks are arranged on the testing platform. When the cylinder to be tested is placed at the target position corresponding to the positioning mark, the plurality of positioning sensors send prompt information to the smart chip indicating that the placement of the cylinder to be tested is complete. Specifically, the positioning sensor is a device for dynamically detecting the cylinder to be tested. The positioning mark is a point of the static reference target position during the positioning process. The positioning sensor needs the positioning mark as a reference to determine the exact position of the cylinder to be tested. A plurality of positioning sensors and positioning marks are arranged on the testing platform. The steps of setting are as follows: first, the positioning sensor is selected according to the specific needs of the cylinder detection. For example, if the precise position of the cylinder needs to be detected on the testing platform, a proximity sensor with high precision is selected. The positioning mark can be a groove that matches the size of the cylinder. A proximity sensor is placed in the groove. When the cylinder to be tested is placed on the groove, the cylinder to be tested contacts the proximity sensor. The proximity sensor sends a prompt information to the smart chip indicating that the placement of the cylinder to be tested is complete. The prompt information can be a prompt signal.
[0035] The smart chip 103 is used to control the second mechanical arm to drive the label reading device to read the cylinder information in the electronic label set on the cylinder to be detected after receiving the prompt information, and parse the cylinder information to obtain multiple attribute information and historical maintenance information of the cylinder to be detected. In some embodiments, after receiving the prompt information, the smart chip controls the second mechanical arm to drive the label reading device (RFID reader) to read the cylinder information in the electronic label. The multiple attribute information includes production date, production batch, and cylinder type. The historical maintenance information includes multiple maintenance records, and each maintenance record includes maintenance time, maintenance location, and maintenance method. In practice, the position information of the electronic label on the cylinder to be detected can be input into the cylinder inner wall processing robot by a technician, so that the smart chip can obtain the position information of the electronic label on the cylinder to be detected and generate a control instruction containing the above position information and the target position, and then the smart chip can send the above control instruction to the second mechanical arm, so that the second mechanical arm drives the label reading device to read the cylinder information in the electronic label.
[0036] The smart chip 103 is also used to determine the service life of the steel cylinder to be tested based on the production date. In some embodiments, assuming that the production date of the steel cylinder to be tested is 2023 and the current time is 2024, the service life of the steel cylinder to be tested is 1 year, and multiple inspection items corresponding to the steel cylinder to be tested are queried in a pre-configured inspection item query table based on the service life. The inspection item query table includes different service lives and corresponding multiple inspection items. The inspection items include but are not limited to: inner wall corrosion detection, inner wall impurity detection, inner wall layer detection and inner wall structural integrity detection. The multiple inspection items queried constitute a candidate inspection item set.
[0037] According to the production batch, multiple inner wall defects corresponding to the steel cylinders of the production batch and the historical number of detections corresponding to each inner wall defect are queried in the defect detection database. In some embodiments, the production batch is a batch of steel cylinders (100) produced in October 2023, and the inner wall defects include damage defects and corrosion defects. The historical number of detections is the historical number of detections corresponding to each inner wall defect in the steel cylinders of the same production batch; from the multiple inner wall defects, the inner wall defects with a historical number of detections greater than the preset number of detections are selected as the inner wall defects to be checked, and the detection items corresponding to the inner wall defects to be checked are added to the candidate detection item set A set of standard inspection items is obtained. In some embodiments, assuming that the preset number of detections is 10 times, the historical number of detections of inner wall defect A is 15 times, and the historical number of detections of inner wall defect B is 8 times, then the inner wall defect A greater than the preset number of detections is taken as the inner wall defect to be checked, and the inner wall defect item inspection table is queried to determine the inspection item corresponding to the inner wall defect to be checked. The inner wall defect item inspection table includes different inner wall defects and corresponding inspection items. If the inner wall defect to be checked is a corrosion defect, the corresponding inspection item may be inner wall corrosion inspection. The inner wall defect B does not meet the conditions and is not the focus of the current inspection.
[0038] The smart chip 103 is also used to generate an inner wall detection plan based on a set of standard detection items. The inner wall detection plan includes image acquisition indicators. According to the inner wall detection plan, the first robotic arm is controlled to drive the image acquisition device to capture the inner wall of the cylinder to be inspected, and an inner wall image set is obtained. The inner wall image set is used to process the inner wall of the cylinder to be inspected.
[0039] In some embodiments, a plurality of positioning sensors and positioning marks are provided on the inspection table, which improves the accuracy of placing the cylinders to be inspected at the target position and enhances the inspection efficiency; in addition, the smart chip queries the plurality of inspection items corresponding to the cylinders to be inspected in a pre-configured inspection item query table according to the service life, and unnecessary inspections of cylinders that have not reached the inspection cycle can be avoided by using the service life, thereby ensuring the targeted inspection and saving inspection resources and costs; according to the production batch, the plurality of inner wall defects corresponding to the cylinders of the production batch and the historical number of detections corresponding to each inner wall defect can be queried in the defect inspection database, so as to quickly understand the inner wall defects of the cylinders of the production batch, conduct targeted inspections and improve the inspection efficiency; according to the set of standard inspection items, an inner wall inspection plan is generated, and by following the standard inspection items, the scientific nature of the generated inner wall inspection plan can be ensured.
[0040] In some embodiments, in order to further solve the second technical problem described in the background technology part, that is, "in the process of collecting images of the inner wall of the steel cylinder, due to the inaccurate shooting direction, the actual condition of the inner wall of the steel cylinder cannot be fully reflected, and the low-resolution image of the key area may increase the difficulty of maintenance and inspection", in some embodiments of the present invention, the image collection index includes the longitudinal shooting height and the lateral rotation angle, and the longitudinal shooting height and the lateral rotation angle are generated by the following steps:
[0041] Step 1: Match the cylinder type in the pre-configured cylinder partitioning scheme library to obtain the cylinder partitioning scheme corresponding to the cylinder to be detected, and the cylinder partitioning scheme includes the number of axial partitions and the number of lateral partitions;
[0042] Step 2: According to the number of axial partitions, the cylinder to be inspected is axially divided to obtain multiple axial partitions, and the axial height interval corresponding to each axial partition is determined;
[0043] Step 3: According to the number of transverse partitions, each of the multiple axial partitions is transversely divided to obtain multiple transverse partitions corresponding to each axial partition, and the rotation angle interval corresponding to each transverse partition is determined. The transverse partitions corresponding to the multiple axial partitions respectively constitute a transverse partition set;
[0044] Step 4: Determine the image acquisition index corresponding to each horizontal partition in the horizontal partition set according to the axial height interval corresponding to each axial partition and the rotation angle interval corresponding to each horizontal partition. The image acquisition index includes the longitudinal shooting height and the horizontal rotation angle.
[0045] Image acquisition metrics include image acquisition resolution, which is generated through the following steps:
[0046] Step 1: Generate a re-inspection score corresponding to each maintenance record in multiple maintenance records according to the maintenance time and maintenance method. If the re-inspection score is greater than a preset re-inspection score threshold, determine the maintenance location contained in the corresponding maintenance record as the target maintenance location, and match the target maintenance location in the horizontal partition set to obtain a horizontal partition matching the target maintenance location and use it as the key horizontal partition;
[0047] Step 2: configure a first image acquisition resolution for the key horizontal partition in the horizontal partition set, and configure a second image acquisition resolution for the remaining horizontal partitions in the horizontal partition set except the key horizontal partition, wherein the first image acquisition resolution is greater than the second image acquisition resolution.
[0048] The inner wall detection scheme is generated through the following steps:
[0049] Step 1: Match the deep learning model corresponding to each detection item from the pre-configured deep learning model library to obtain a deep learning model set corresponding to the standard detection item set;
[0050] Step 2: Combine the deep learning model set and image acquisition indicators into an inner wall detection solution.
[0051] Optionally, a steel cylinder inner wall processing robot of the present invention further includes a display device; the smart chip is also used to sort the various standard detection items in the standard detection item set according to the order of project execution to obtain a standard detection item sequence; the standard detection item sequence and the current execution state of each standard detection item in the standard detection item sequence are sent to the display device; the display device is used to receive the standard detection item sequence and the current execution state of each standard detection item and display them, and when receiving the user's adjustment information for the displayed standard detection item sequence, the adjustment information is sent to the smart chip, so that the smart chip sequentially adjusts the multiple standard detection items in the standard detection item sequence whose current execution state is waiting for execution according to the adjustment information, obtains a new standard detection item sequence, and updates the inner wall detection scheme according to the new standard detection item sequence. Thus, the execution order between the various standard detection items can be included in the inner wall detection scheme.
[0052] In some embodiments, the longitudinal shooting height and the lateral rotation angle can accurately capture specific parts of the cylinder and can fully capture any area of the cylinder, thereby improving the efficiency of subsequent image analysis; a higher image acquisition resolution is configured for key target maintenance locations so that the collected images can provide more information, thereby obtaining a more scientific and accurate inner wall detection solution.
[0053] In some embodiments, in order to further solve the technical problem three described in the background technology section, that is, "the traditional method for extracting texture feature data of the inner wall of the steel cylinder has a low degree of automation, it is difficult to extract effective texture feature data from the local area, and the obtained texture feature result representation method is inefficient, which causes difficulties in subsequent processing, and it is impossible to accurately identify all types of inner wall defects, and it is impossible to remove the rust on the inner wall of the steel cylinder in time", in some embodiments of the present invention, the second mechanical arm of the two mechanical arms is equipped with a rust removal brush, and the smart chip is also used for:
[0054] Convert each inner wall image in the inner wall image set into a grayscale image to obtain a grayscale image set; divide each grayscale image in the grayscale image set into a plurality of image blocks, and the size of each image block is a preset image size;
[0055] Determine a central pixel of each image block in the plurality of image blocks, determine pixels within a preset range of the central pixel as a neighborhood pixel group, and sort the neighborhood pixel groups in a clockwise order to obtain a neighborhood pixel sequence;
[0056] If the gray value of the neighboring pixel is greater than the gray value of the central pixel, the comparison result corresponding to the neighboring pixel is set to 1; if the gray value of the neighboring pixel is less than or equal to the gray value of the central pixel, the comparison result corresponding to the neighboring pixel is set to 0;
[0057] According to the neighborhood pixel sequence, the comparison results corresponding to each neighborhood pixel are spliced to obtain a comparison result sequence, the comparison result sequence is encoded to obtain a sequence code, and the sequence code is determined as a texture feature of each image block; the texture features corresponding to multiple image blocks are spliced according to the positions of the image blocks in the grayscale image to obtain an image texture feature corresponding to each grayscale image;
[0058] Input the image texture features into a pre-trained cylinder inner wall detection network to obtain the type of inner wall defect; in some embodiments, the cylinder inner wall detection network (such as a softmax layer) is used to detect various inner wall defect types of the cylinder inner wall, and the inner wall defect type may be rust or corrosion;
[0059] If the type of inner wall defect is rust, a rust removal instruction is generated and transmitted to the second robotic arm through the data transmission line, and the second robotic arm drives the rust removal brush to perform rust removal operations. Among them, the cylinder inner wall detection network can also output the location information of the inner wall defect, so that the rust removal instruction can include the location information of the inner wall defect, so that the second robotic arm drives the rust removal brush to perform rust removal operations at the corresponding position.
[0060] In some embodiments, the central pixel can be regarded as a two-dimensional coordinate point (x, y), where x and y represent the column and row indices of the pixel in the image, respectively; the neighborhood pixel group is the pixel points within a certain range around the central pixel. As an example, in a matrix of a preset image size of 3x3, the central pixel is located at the center of the neighborhood, and the surrounding 8 pixels constitute the neighborhood pixel group.
[0061] In some embodiments, the grayscale value of each neighborhood pixel and the grayscale value of the corresponding center pixel are directly obtained from a grayscale image. In a grayscale image, each pixel is represented by an integer between 0 (black) and 255 (white), which is the grayscale value of the pixel.
[0062] In some embodiments, if the preset image size is a 3x3 matrix, assuming that the grayscale value of the central pixel is 6, the grayscale values of the neighborhood pixel group are (in clockwise order): 6, 5, 2, 1, 7, 3, 9, 7; the grayscale value of each neighborhood pixel in the neighborhood pixel group is compared with the grayscale value of the corresponding central pixel: 6 is compared with 6, the grayscale value of the neighborhood pixel is equal to the grayscale value of the central pixel, and the comparison result corresponding to each neighborhood pixel is set to 0; 6 is compared with 5, and the grayscale value of the neighborhood pixel is greater than the grayscale value of the central pixel, then the comparison result corresponding to each neighborhood pixel is set to 1; neighborhood pixels 2, 1, 7, 3, 9, 7 and so on are used to obtain the comparison result corresponding to each neighborhood pixel; the comparison results corresponding to each neighborhood pixel are spliced to obtain a comparison result sequence: 01001110, and the comparison result sequence: 01001110 is encoded to obtain a sequence code (converted to decimal): 78, and 78 is determined as texture feature data.
[0063] In some embodiments, during the extraction of texture feature data by the smart chip, the efficiency of the local processing process is improved by dividing the image into multiple image blocks; by comparing the neighborhood pixels with the central pixels, the changes in local texture features can be captured, and the encoding obtained by comparing the domain pixels with the central pixels is a binary number, which provides an efficient texture feature representation method, facilitates the subsequent texture feature recognition of the image, and the calculation is simple, efficient and highly automated. The pre-trained network can quickly process the input texture feature data and quickly identify the type of texture features. Timely rust removal can avoid potential safety risks caused by rust.
[0064] The above descriptions are only some preferred embodiments of the present invention and the explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, but also should cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present invention (but not limited to) to form a technical solution.
Claims
1. A robot for processing the inner wall of a steel cylinder, characterized in that: include: Two mechanical arms, the two mechanical arms cooperate with each other to place the steel cylinder to be tested on the testing table, the end of the first mechanical arm of the two mechanical arms is provided with an image acquisition device, and the end of the second mechanical arm of the two mechanical arms is provided with a label reading device; A testing platform, the testing platform is used to place the steel cylinder to be tested, and a plurality of positioning sensors and positioning marks are arranged on the testing platform. When the steel cylinder to be tested is placed at a target position corresponding to the positioning mark, the plurality of positioning sensors send prompt information indicating that the placement of the steel cylinder to be tested is completed to the smart chip; A smart chip, wherein after receiving the prompt information, the smart chip is used to control the second mechanical arm to drive the label reading device to read the cylinder information in the electronic label set on the cylinder to be detected, and parse the cylinder information to obtain multiple attribute information and historical maintenance information of the cylinder to be detected, wherein the multiple attribute information includes production date, production batch, and cylinder type, and the historical maintenance information includes multiple maintenance records, each maintenance record includes maintenance time, maintenance location, and maintenance method; The smart chip is also used to determine the service life of the steel cylinder to be tested according to the production date, and query multiple test items corresponding to the steel cylinder to be tested in a pre-configured test item query table according to the service life, and the multiple test items queried constitute a candidate test item set; query multiple inner wall defects corresponding to the steel cylinders of the production batch and the historical detection times corresponding to each inner wall defect in a defect detection database according to the production batch; select the inner wall defect whose historical detection times are greater than the preset detection times from the multiple inner wall defects as the inner wall defect to be checked, and add the test items corresponding to the inner wall defect to be checked to the candidate test item set to obtain a standard test item set; The smart chip is also used to generate an inner wall detection plan based on the set of standard detection items, and the inner wall detection plan includes image acquisition indicators. According to the inner wall detection plan, the first robotic arm is controlled to drive the image acquisition device to capture the image of the inner wall of the steel cylinder to be inspected, and an inner wall image set is obtained. The inner wall image set is used to process the inner wall of the steel cylinder to be inspected.
2. The cylinder inner wall processing robot according to claim 1, characterized in that: The image acquisition index includes a vertical shooting height and a horizontal rotation angle, and the vertical shooting height and the horizontal rotation angle are generated by the following steps: Matching the cylinder type in a pre-configured cylinder partitioning scheme library to obtain a cylinder partitioning scheme corresponding to the cylinder to be detected, wherein the cylinder partitioning scheme includes the number of axial partitions and the number of lateral partitions; According to the number of axial partitions, the steel cylinder to be inspected is axially divided to obtain a plurality of axial partitions, and an axial height interval corresponding to each axial partition is determined; According to the number of transverse partitions, each of the multiple axial partitions is transversely divided to obtain multiple transverse partitions corresponding to each axial partition, and a rotation angle interval corresponding to each transverse partition is determined, and the transverse partitions corresponding to the multiple axial partitions respectively constitute a transverse partition set; According to the axial height interval corresponding to each axial partition and the rotation angle interval corresponding to each lateral partition, the image acquisition index corresponding to each lateral partition in the lateral partition set is determined, and the image acquisition index includes the longitudinal shooting height and the lateral rotation angle.
3. The cylinder inner wall processing robot according to claim 2, characterized in that: The image acquisition index includes an image acquisition resolution, and the image acquisition resolution is generated by the following steps: Generate a re-inspection score corresponding to each maintenance record in the multiple maintenance records according to the maintenance time and the maintenance method; if the re-inspection score is greater than a preset re-inspection score threshold, determine the maintenance location contained in the corresponding maintenance record as a target maintenance location, and match the target maintenance location with the horizontal partition set to obtain a horizontal partition matching the target maintenance location and use it as a key horizontal partition; A first image acquisition resolution is configured for a key horizontal partition in the horizontal partition set, and a second image acquisition resolution is configured for the remaining horizontal partitions in the horizontal partition set except the key horizontal partition, wherein the first image acquisition resolution is greater than the second image acquisition resolution.
4. The cylinder inner wall processing robot according to claim 3, characterized in that: The inner wall detection scheme is generated by the following steps: Matching the deep learning model corresponding to each detection item from a pre-configured deep learning model library to obtain a deep learning model set corresponding to the standard detection item set; The deep learning model set and the image acquisition index form the inner wall detection solution.
5. The cylinder inner wall processing robot according to claim 4, characterized in that: Also includes a display device; and The smart chip is also used to sort the various standard test items in the standard test item set according to the order of project execution to obtain a standard test item sequence; the standard test item sequence and the current execution status of each standard test item in the standard test item sequence are sent to a display device; The display device is used to receive and display the standard detection item sequence and the current execution status of each standard detection item. When receiving adjustment information from the user for the displayed standard detection item sequence, the display device sends the adjustment information to the smart chip, so that the smart chip sequentially adjusts multiple standard detection items in the standard detection item sequence whose current execution status is waiting for execution according to the adjustment information, obtains a new standard detection item sequence, and updates the inner wall detection plan according to the new standard detection item sequence.
Citation Information
Patent Citations
Evaluation method of wine bottle surface defect model, electronic device and storage medium
CN111709452A
Multi-scale ceramic detection method and system
CN113237889A
Image acquisition equipment and quality detection system
CN116858848A
Intelligent part detection device and method based on industrial vision technology
CN117990705A
Bottle wall defect detection method and system for high-pressure gas cylinder
CN118032794A
