A laser scanning-based method and system for detecting screw residue
By acquiring three-dimensional point cloud data of the bottom of the basketball hoop through laser scanning and combining it with multi-dimensional feature quantification, the problem of automation and consistency in screw residue detection in fastener production has been solved, achieving efficient and accurate fully automated detection throughout the entire process.
Patent Information
- Application Number
- CN202611081357.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-21
- Publication Date
- 2026-11-10
AI Technical Summary
In the fastener production process, existing technologies are unable to achieve full-area scanning and highly interference-resistant fully automated screw residue detection, resulting in inconsistent detection results and low efficiency, which affects production efficiency and cost.
A laser scanning-based method is used to acquire three-dimensional point cloud data of the bottom of the basketball hoop. This data is then combined with differential comparison of baseline data and multi-dimensional feature quantification. Finally, an audible and visual alarm and a diversion mechanism are used to automate the entire process of detection, early warning, and diversion.
It enables accurate identification of residual fasteners in blind spots and gaps at the bottom of the basketball hoop, reduces false alarm rate, ensures consistency of test results, automates the entire process, and reduces labor and scrap costs.
Smart Images

Figure CN123066876A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of fastener manufacturing technology, specifically relating to a method and system for detecting screw residue based on laser scanning. Background Technology
[0002] In the mass production of fasteners (such as screws and bolts), turnover baskets are typically used for inter-process handling and unloading. In actual production, small screws easily get stuck in the gaps at the bottom of the basket, the corners of reinforcing ribs, etc., during unloading, often leaving a small amount of residue inside the basket, which is difficult to completely empty. If this residue is not detected and removed promptly, when the empty basket is returned to the front of the production line for reloading, screws from different batches and of different specifications will be mixed with the new material, causing batch mixing, resulting in the scrapping of the entire batch of products, leading to serious quality incidents and customer complaints.
[0003] Therefore, strict residue testing is essential before empty baskets return to the production line. Currently, the most common method in the industry for this step is manual visual inspection. However, the results are highly dependent on the worker's experience, eyesight, and mental state, making it impossible to guarantee consistent judgment standards across batches. Furthermore, manual inspection is inefficient and difficult to match the pace of automated production lines, impacting production efficiency and costs. Some companies commonly use CCD cameras for detection, but 2D visual inspection is highly dependent on consistent lighting. Changes in natural light in the workshop environment, interference from light sources from adjacent equipment, and alterations in reflectivity due to oil and dust on the basket surface all directly affect the grayscale distribution of the image, resulting in significant differences in the imaging characteristics of the same object at different times, easily leading to false alarms or missed detections. Moreover, 2D visual inspection cannot detect the height information of objects. When a screw remains deep in a gap, it can only be seen as a tiny dot in a 2D image, indistinguishable from the inherent texture of the bottom groove. Additionally, during use, the baskets inevitably develop surface scratches and oil residue. These defects, in 2D images, exhibit grayscale features highly similar to screw residue, making it difficult for traditional image processing algorithms to effectively distinguish them. This results in numerous false alarms, severely impacting the normal operation efficiency of the production line.
[0004] Therefore, there is an urgent need for a fully automated screw residue detection solution that can achieve full-area scanning and has strong anti-interference capabilities. Summary of the Invention
[0005] Purpose of the invention: To overcome the above shortcomings, this invention provides a method and system for detecting screw residue based on laser scanning. It acquires three-dimensional point cloud data of the entire bottom of the basketball hoop through two-dimensional line laser scanning, and combines the differential comparison of benchmark data with multi-dimensional feature quantification to achieve accurate identification of residual fasteners in dead corners and gaps at the bottom of the basketball hoop. Furthermore, it integrates audible and visual alarms and a diversion mechanism to automate the entire process of detection, early warning, and diversion, avoiding material mixing problems caused by residual materials and reducing quality inspection and scrapping costs.
[0006] Technical solution: In order to achieve the above objectives, the present invention provides a screw residue detection method based on laser scanning, including the following steps: S1: Reference modeling: Obtain the three-dimensional point cloud data of the entire bottom area of a standard unloaded turnover basket as reference contour data; S2: Online scanning: Laser scanning is performed on the entire bottom area of the test basket to obtain actual 3D point cloud data; S3: Differential preprocessing: Perform height difference calculation on the measured point cloud data and the reference point cloud data point by point according to spatial coordinates to obtain a height difference distribution map; S4: Filtering and noise reduction: The height difference distribution map is filtered to eliminate noise interference; S5: Foreground segmentation: Based on the comparison between the filtered height difference and the preset height threshold, extract the height protrusion area as the candidate residual material area; S6: Connected component filtering: Perform connected component analysis on the candidate residual material region and remove isolated noise regions with an area smaller than a preset area threshold. S7: Feature Quantization: Calculate multi-dimensional quantization features for each candidate residual material region after screening. The multi-dimensional quantization features include the average bulge height of the region, the ratio of the major and minor axes of the region, and the area of the region. S8: Confidence determination: Based on the multidimensional quantitative features of each candidate residual material area, calculate the matching confidence of each area with the standard fastener features, and determine the overall confidence of the whole basket based on the confidence of each area. Compare the overall confidence of the whole basket with the preset qualified judgment threshold and output the judgment result. S9: Output signal: Output the corresponding control signal according to the judgment result.
[0007] Furthermore, in the above-described laser scanning-based screw residue detection method, the height difference calculation in step S3 is performed according to the following formula: in, This represents the pixel position coordinates arranged along the scan width direction and along the transport direction in the laser scanning data. For row index, For column indexes; The coordinate position in the measured 3D point cloud data The height value at that location, For the same coordinate position in the reference contour data The height value at that location, coordinate position The height difference at each location. The height value is positive when vertically upward. By subtracting the baseline height of the empty basket pixel by pixel, the inherent height fluctuations of the basket's bottom surface and gaps are effectively eliminated, retaining only the height increase caused by foreign objects.
[0008] The filtering process described in step S4 uses Gaussian filtering, and is performed according to the following formula: in, The filtered coordinate position The height difference at that location Coordinates before filtering The height difference at that location The Gaussian kernel function at the offset The value at that location, This represents the offset along the row direction. Offset along the column direction, Let the radius be the Gaussian kernel. The standard deviation of the Gaussian kernel is used. Gaussian filtering is applied to smooth the height difference distribution map, suppressing sensor measurement noise and interference caused by the microscopic roughness of the basket surface.
[0009] Furthermore, in the above-described laser scanning-based screw residue detection method, the foreground segmentation step in step S5 is performed according to the following formula: in, coordinate position ( The binarization marker value at () indicates that the pixel is a foreground pixel, and 0 indicates that the pixel is a background pixel; The preset height threshold is determined by calibration based on the surface roughness of the basket and the measurement noise of the scanning module.
[0010] In step S6, the connected component analysis includes, for For the foreground pixel with a value of 1, perform 8-neighbor connected component extraction to obtain K independent connected components. , , ..., Calculate the area of each connected component using the following formula: in, For the first The area of each connected region; area Less than the preset area threshold Connected component removal, retained area ≥ The connected components constitute the set of effective residual material regions M={ | ≥ The area is smaller than the area of the first k, k=1,2,…,K. Tiny, isolated areas are considered noise points (such as dust, metal shavings, etc.) and are directly removed.
[0011] Furthermore, in the above-mentioned laser scanning-based screw residue detection method, in the multi-dimensional quantification features described in step S7: the first... The average protrusion height of each effective residual material area is calculated using the following formula: No. The ratio of the major and minor axes of each effective residual material region is calculated using the following formula: in, For the first The major axis dimension of the minimum bounding rectangle of the effective residual material area. For the first The minor axis dimension of the smallest bounding rectangle of an effective residual material area.
[0012] Furthermore, in the above-mentioned laser scanning-based screw residue detection method, in step S8, the matching confidence level between each effective residue area and the standard fastener features is calculated according to the following formula: in, The average protrusion height of standard residual fasteners. This represents the average ratio of the major and minor axes of a standard fastener. The average pixel area for standard fastener imaging; , , These represent the standard deviations of height, proportion, and area features, respectively. The matching scores are calculated using a Gaussian kernel function, with a range of (0,1]. The closer the feature value is to the mean feature value of the calibrated fastener, the closer the matching score is to 1.
[0013] No. The confidence level of a single region for each effective residual material region is calculated using the following formula: in, , , The weights are respectively for height feature, shape proportion feature, and area feature, and satisfy the following conditions: Each weight is assigned a value through calibration based on actual process requirements.
[0014] Furthermore, in the above-mentioned laser scanning-based screw residue detection method, the overall confidence level of the entire basket in step S8 is calculated according to the following formula: The maximum confidence level of a single region among all valid residual material regions is taken as the overall confidence level of the entire basket. For the collection of effective residual material areas, The first in the set of effective residual material regions Connected components, For the first The confidence level of a single region in a connected component.
[0015] The determination logic is as follows: when If a residual fastener is found inside the test basket, an unqualified signal is output. when When it is determined that there are no residual fasteners inside the test basket, a pass signal is output. When the effective residual material area is collected When the set is empty, it is directly determined that there are no residual fasteners in the test basket and a qualified signal is output. in, The preset pass / fail threshold is defined as follows.
[0016] Furthermore, in the aforementioned laser scanning-based screw residue detection method, when it is determined that residual fasteners exist within the test basket, the physical center coordinates of the residue area are calculated according to the following formula: in, coordinate position The corresponding physical coordinates of the conveying direction coordinate position The corresponding physical coordinates in the width direction; and The first The center coordinates of each residual material area within the basket are used to coordinate the sorting mechanism for positioning and isolation.
[0017] Furthermore, in the above-mentioned laser scanning-based screw residue detection method, in the benchmark modeling step, multiple sets of benchmark contour data corresponding to various specifications of turnover baskets are pre-stored, and the corresponding benchmark contour data are retrieved according to the specifications of the basket to be tested before detection.
[0018] Furthermore, the aforementioned laser scanning-based screw residue detection method also includes data traceability: the judgment results of each test, the physical center coordinates of the residue area, and the confidence level of each area are recorded and uploaded to the production management system to achieve traceability of the test data.
[0019] A laser scanning-based screw residue detection system for turnover baskets includes: a conveyor line for carrying and transporting turnover baskets; a laser scanning module mounted above the conveyor line via a bracket for performing laser scanning on the entire bottom area of the basket and acquiring measured three-dimensional point cloud data; a storage unit for pre-storing reference three-dimensional point cloud data of the entire bottom area of a standard unloaded turnover basket; a data processing unit connected to the two-dimensional laser scanning module and the storage unit, wherein the data processing unit executes the aforementioned detection method and outputs a residue judgment signal; an audible and visual alarm device connected to the data processing unit for issuing an audible and visual alarm when the residue judgment signal is unqualified; and a diversion mechanism located downstream of the conveyor line and connected to the data processing unit for guiding turnover baskets containing residual fasteners to a cleaning area according to the residue judgment signal.
[0020] As can be seen from the above technical solution, the present invention has the following beneficial effects: ①The present invention is a screw residue detection method and system based on laser scanning. It uses a two-dimensional line laser scanning module to perform a full-area scan on the bottom of the turnover basket. The laser line covers the width direction of the basket, and the uniform movement of the conveyor line covers the length direction, thus synthesizing a surface scan that covers the entire bottom of the basket, achieving full coverage of the bottom of the basket and eliminating detection blind spots.
[0021] ② This invention is not affected by environmental factors such as oil stains on the surface of the basket, color changes, and changes in light intensity, and has good environmental adaptability.
[0022] ③ This invention uses multi-dimensional feature fusion confidence calculation based on high matching score, shape ratio matching score, and area matching score to effectively distinguish fasteners from other debris such as dust and metal shavings. While ensuring a high detection rate, it significantly reduces the false alarm rate and achieves full quantitative judgment of residual fasteners.
[0023] ④ This invention integrates online detection, audible and visual alarms, automatic diversion, and data traceability functions, realizing fully automated operation from detection to disposal without manual intervention, eliminating the risk of residual materials being mixed into the next batch, and reducing labor costs and scrap costs. Attached Figure Description
[0024] Figure 1 The flowchart of the screw residue detection method based on laser scanning of the present invention is as follows; Figure 2 This is a schematic diagram of the structure of the laser scanning-based screw residue detection system for turnover baskets.
[0025] In the diagram: 1. Conveyor line, 3. Turnover basket, 2. Laser scanning module, 4. Data processing unit, 5. Audible and visual alarm device. Detailed Implementation
[0026] Example 1 like Figure 1 The system described is a laser scanning-based screw residue detection system for turnover baskets, comprising: a conveyor line 1 for carrying and transporting turnover baskets 3 (in this embodiment, the conveyor line 1 is a roller conveyor line); a laser scanning module 2 mounted above the conveyor line 1 via a bracket for performing laser scanning on the entire bottom area of the basket 3 and acquiring measured three-dimensional point cloud data; a storage unit for pre-storing reference three-dimensional point cloud data of the entire bottom area of a standard unloaded turnover basket 3; a data processing unit 4 connected to the two-dimensional laser scanning module and the storage unit, wherein the data processing unit 4 executes the aforementioned detection method and outputs a residue judgment signal; an audible and visual alarm device 5 connected to the data processing unit 4 for issuing an audible and visual alarm when the residue judgment signal is unqualified; and a diversion mechanism located downstream of the conveyor line and connected to the data processing unit for guiding the turnover basket 3 containing residual fasteners to a cleaning area according to the residue judgment signal. The conveyor line 1 has a positioning mechanism at the scanning station for positioning the basket 2 directly below the laser scanning module 3. The conveyor line 1 is also equipped with an encoder 11, which is used to collect the displacement pulse signal of the basket 2 in real time and send it to the data processing unit 5 to realize the spatial coordinate alignment between the measured three-dimensional point cloud data and the reference contour data.
[0027] In this embodiment, the laser scanning module 3 uses a two-dimensional line laser profile measuring instrument. The direction of the laser line it projects covers the width direction of the basket 2. When the basket 2 is driven by the conveyor line 1 to pass under the scanning module 3 at a constant speed along the length direction, a scanning area covering the entire bottom area of the basket 2 is synthesized.
[0028] Example 2 like Figure 2The method for detecting screw residue based on laser scanning includes the following steps: S1: Baseline modeling: Acquire three-dimensional point cloud data of the entire bottom area of a standard empty turnkey basketball hoop as baseline contour data; pre-store multiple sets of baseline contour data for turnkey basketball hoops of various specifications, and retrieve the corresponding baseline contour data according to the specifications of the basketball hoop to be tested before detection. Specifically, take a standard empty basketball hoop that has been thoroughly cleaned and confirmed to be free of any residue, scan it 5 times consecutively, and take the arithmetic mean of the 5 height values at the same spatial coordinate point after removing the maximum and minimum values to generate the baseline contour data of the basketball hoop.
[0029] S2: Online scanning: Laser scanning is performed on the entire bottom area of the test basket to obtain actual 3D point cloud data; S3: Differential preprocessing: Perform height difference calculation on the measured point cloud data and the reference point cloud data point by point according to spatial coordinates to obtain a height difference distribution map; The height difference calculation is performed according to the following formula: in, This represents the pixel position coordinates arranged along the scan width direction and along the transport direction in the laser scanning data. For row index, For column indexes; The coordinate position in the measured 3D point cloud data The height value at that location, For the same coordinate position in the reference contour data The height value at that location, coordinate position The height difference at each location. The height value is positive when vertically upward. By subtracting the baseline height of the empty basket pixel by pixel, the inherent height fluctuations of the basket's bottom surface and gaps are effectively eliminated, retaining only the height increase caused by foreign objects.
[0030] S4: Filtering and noise reduction: The height difference distribution map is filtered to eliminate noise interference; The filtering process uses Gaussian filtering, and is performed according to the following formula: in, The filtered coordinate position The height difference at that location Coordinates before filtering The height difference at that location The Gaussian kernel function at the offset The value at that location, This represents the offset along the row direction. Offset along the column direction, Let the radius be the Gaussian kernel. The standard deviation of the Gaussian kernel is used. Gaussian filtering is applied to smooth the height difference distribution map, suppressing sensor measurement noise and interference caused by the microscopic roughness of the basket surface.
[0031] S5: Foreground segmentation: Based on the comparison between the filtered height difference and the preset height threshold, extract the height protrusion area as the candidate residual material area; The foreground segmentation step is performed according to the following formula: in, coordinate position ( The binarization marker value at () indicates that the pixel is a foreground pixel, and 0 indicates that the pixel is a background pixel; The preset height threshold is determined through calibration based on the surface roughness of the basketball hoop and the measurement noise of the scanning module. In this embodiment, The height distinction threshold is 0.1~0.5mm, determined based on the surface roughness of the turnover basket and the noise level of the laser scanning module 3.
[0032] S6: Connected component filtering: Perform connected component analysis on the candidate residual material region and remove isolated noise regions with an area smaller than a preset area threshold. Connected component analysis includes, for For the foreground pixel with a value of 1, perform 8-neighbor connected component extraction to obtain K independent connected components. , …, Calculate the area of each connected component using the following formula: in, For the first The area of each connected region; area Less than the preset area threshold Connected component removal based on the pixel area threshold for filtering out minute noise, retaining the area. ≥ The connected components constitute the set of effective residual material regions M={ | ≥ The area is smaller than the area of the first k, k=1,2,…,K. Tiny, isolated areas are considered noise points (such as dust, metal shavings, etc.) and are directly removed. In this embodiment, The resolution is 40 to 100 pixels, determined based on the minimum imaging pixel area of the fastener to be tested.
[0033] S7: Feature Quantization: Calculate multi-dimensional quantization features for each candidate residual material region after screening. The multi-dimensional quantization features include the average bulge height of the region, the ratio of the major and minor axes of the region, and the area of the region. In the aforementioned multidimensional quantization features: the first The average protrusion height of each effective residual material area is calculated using the following formula: No. The ratio of the major and minor axes of each effective residual material region is calculated using the following formula: in, For the first The major axis dimension of the minimum bounding rectangle of the effective residual material area. For the first The minor axis dimension of the smallest bounding rectangle of an effective residual material area.
[0034] The area is calculated in step S6. .
[0035] S8: Confidence determination: Based on the multidimensional quantitative features of each candidate residual material area, calculate the matching confidence of each area with the standard fastener features, and determine the overall confidence of the whole basket based on the confidence of each area. Compare the overall confidence of the whole basket with the preset qualified judgment threshold and output the judgment result. The confidence level of matching between each effective residual material area and the characteristics of the standard fastener is calculated according to the following formula: in, The average protrusion height of standard residual fasteners. This represents the average ratio of the major and minor axes of a standard fastener. The average pixel area for standard fastener imaging; , , These represent the sample standard deviations corresponding to the three types of features. All parameters were obtained through statistical calibration of multiple sets of standard fastener point cloud samples. The matching score was calculated using a Gaussian kernel function, with a value range of (0,1]. The closer the feature value is to the mean feature value of the calibrated fastener, the closer the matching score is to 1.
[0036] in, , , The weights are respectively for height feature, shape proportion feature, and area feature, and satisfy the following conditions: Each weight is assigned a value through calibration based on actual process requirements.
[0037] The overall confidence level is calculated using the following formula: The maximum confidence level of a single region among all valid residual material regions is taken as the overall confidence level of the entire basket. For the collection of effective residual material areas, The first in the set of effective residual material regions Connected components, For the first The confidence level of a single region in a connected component.
[0038] The determination logic is as follows: when If a residual fastener is found inside the test basket, an unqualified signal is output. when When it is determined that there are no residual fasteners inside the test basket, a pass signal is output. When the effective residual material area is collected When the set is empty, it is directly determined that there are no residual fasteners in the test basket and a qualified signal is output. in, The preset pass / fail threshold is used. In this embodiment, The value is 0.5-0.8, calibrated according to the equipment's need to balance missed and false detections.
[0039] When it is determined that there are residual fasteners inside the test basket, the physical center coordinates of the residual material area are calculated according to the following formula: in, coordinate position The corresponding physical coordinates of the conveying direction coordinate position The corresponding physical coordinates in the width direction; and The first The center coordinates of each residual material area within the basket are used to coordinate the sorting mechanism for positioning and isolation.
[0040] S9: Output Signal: Outputs the corresponding control signal based on the judgment result. If it fails, the audible and visual alarm device and the diversion mechanism are triggered simultaneously; if it passes, the conveyor line releases the basket to the next process.
[0041] It also includes data traceability: the judgment results of each test, the physical center coordinates of the residual material area and the confidence level of each area are recorded and uploaded to the production management system to achieve traceability of test data.
[0042] The above embodiments are exemplary and are intended to illustrate the technical concept and features of the present invention, so that those skilled in the art can understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made according to the spirit and essence of the present invention should be covered within the scope of protection of the present invention.
Claims
1. A method for detecting screw residue based on laser scanning, characterized in that, Includes the following steps: S1: Benchmark Modeling: Obtain the three-dimensional point cloud data of the entire bottom area of the standard unloaded turnover basket as benchmark contour data; S2: Online scanning: Laser scanning is performed on the entire bottom area of the test basket to obtain actual 3D point cloud data; S3: Differential preprocessing: Perform height difference calculation on the measured point cloud data and the reference point cloud data point by point according to spatial coordinates to obtain a height difference distribution map; S4: Filtering and noise reduction: The height difference distribution map is filtered to eliminate noise interference; S5: Foreground segmentation: Based on the comparison between the filtered height difference and the preset height threshold, extract the height protrusion area as the candidate residual material area; S6: Connected component filtering: Perform connected component analysis on the candidate residual material region and remove isolated noise regions with an area smaller than a preset area threshold. S7: Feature Quantization: Calculate multi-dimensional quantization features for each candidate residual material region after screening. The multi-dimensional quantization features include the average bulge height of the region, the ratio of the major and minor axes of the region, and the area of the region. S8: Confidence determination: Based on the multidimensional quantitative features of each candidate residual material area, calculate the matching confidence of each area with the standard fastener features, and determine the overall confidence of the whole basket based on the confidence of each area. Compare the overall confidence of the whole basket with the preset qualified judgment threshold and output the judgment result. S9: Output signal: Output the corresponding control signal according to the judgment result.
2. The screw residue detection method based on laser scanning according to claim 1, characterized in that: The height difference calculation in step S3 is performed according to the following formula: in, This represents the pixel position coordinates arranged along the scan width direction and along the transport direction in the laser scanning data. For row index, For column indexes; The coordinate position in the measured 3D point cloud data The height value at that location, For the same coordinate position in the reference contour data The height value at that location, coordinate position The height difference at the location; The filtering process described in step S4 uses Gaussian filtering, and is performed according to the following formula: in, The filtered coordinate position The height difference at that location Coordinates before filtering The height difference at that location The Gaussian kernel function at the offset The value at that location, This represents the offset along the row direction. Offset along the column direction, Let the radius be the Gaussian kernel. denoted as the standard deviation of the Gaussian kernel.
3. The method for detecting screw residue based on laser scanning according to claim 2, characterized in that: The foreground segmentation step in step S5 is performed according to the following formula: in, coordinate position ( The binarization marker value at () indicates that the pixel is a foreground pixel, and 0 indicates that the pixel is a background pixel; The preset height threshold; In step S6, the connected component analysis includes: […]. For the foreground pixel with a value of 1, perform 8-neighbor connected component extraction to obtain K independent connected components. , , ..., Calculate the area of each connected component using the following formula: in, For the first The area of each connected region; area Less than the preset area threshold Connected component removal, retained area ≥ The connected components constitute the set of effective residual material regions M={ | ≥ ,k=1,2,…,K}.
4. The screw residue detection method based on laser scanning according to claim 3, characterized in that: In step S7, the multidimensional quantization features include: the first The average protrusion height of each effective residual material area is calculated using the following formula: No. The ratio of the major and minor axes of each effective residual material region is calculated using the following formula: in, For the first The major axis dimension of the minimum bounding rectangle of the effective residual material area. For the first The minor axis dimension of the smallest bounding rectangle of an effective residual material area.
5. The method for detecting screw residue based on laser scanning according to claim 3, characterized in that: In step S8, the matching confidence level between each effective residual material area and the standard fastener features is calculated according to the following formula: in, The average protrusion height of standard residual fasteners. This represents the average ratio of the major and minor axes of a standard fastener. The average pixel area for standard fastener imaging; , , These are the standard deviations of height characteristics, external proportion characteristics, and area characteristics, respectively. No. The confidence level of a single region for each effective residual material region is calculated using the following formula: in, , , The weights are respectively for height feature, shape proportion feature, and area feature, and satisfy the following conditions: .
6. The screw residue detection method based on laser scanning according to claim 1, characterized in that: The overall confidence level of the whole basket mentioned in step S8 is calculated according to the following formula: The maximum confidence level of a single region among all valid residual material regions is taken as the overall confidence level of the entire basket. For the collection of effective residual material areas, The first in the set of effective residual material regions Connected components, For the first Single-region confidence of each connected component; The determination logic is as follows: when If a residual fastener is found inside the test basket, an unqualified signal is output. when When it is determined that there are no residual fasteners inside the test basket, a pass signal is output. When the effective residual material area is collected When the set is empty, it is directly determined that there are no residual fasteners in the test basket and a qualified signal is output. in, The preset pass / fail threshold is defined as follows.
7. The screw residue detection method based on laser scanning according to claim 1, characterized in that: When it is determined that there are residual fasteners inside the test basket, the physical center coordinates of the residual material area are calculated according to the following formula: in, coordinate position The corresponding physical coordinates of the conveying direction coordinate position The corresponding physical coordinates in the width direction; and The first The center coordinates of each residual material area within the basket are used to coordinate the sorting mechanism for positioning and isolation.
8. The method for detecting screw residue based on laser scanning according to claim 1, characterized in that: In the benchmark modeling step, multiple sets of benchmark contour data corresponding to various specifications of turnover baskets are pre-stored. Before testing, the corresponding benchmark contour data is retrieved according to the specifications of the basket to be tested.
9. The method for detecting screw residue based on laser scanning according to claim 1, characterized in that: It also includes data traceability: the judgment results of each test, the physical center coordinates of the residual material area and the confidence level of each area are recorded and uploaded to the production management system to achieve traceability of test data.
10. A laser scanning-based system for detecting residual screws in a turnover basket, characterized in that, include: Conveyor lines are used to carry and transport turnover baskets; The laser scanning module is mounted above the conveyor line via a bracket and is used to perform laser scanning on the entire bottom area of the basket and collect measured three-dimensional point cloud data. Storage unit, used to pre-store the baseline three-dimensional point cloud data of the entire bottom area of the standard empty turnover basket; A data processing unit is connected to the two-dimensional laser scanning module and the storage unit respectively. The data processing unit executes the detection method as described in any one of claims 1 to 9 and outputs a residual material determination signal. An audible and visual alarm device, connected to the data processing unit, is used to issue an audible and visual alarm when the residual material judgment signal is unqualified. A diversion mechanism, located downstream of the conveyor line and connected to the data processing unit, is used to guide baskets containing residual fasteners to a cleaning area based on the residual material determination signal.
