Performance test method based on data analysis and safety helmet test system

Abstract

The invention relates to the technical field of performance testing, and discloses a performance testing method based on data analysis and a safety helmet testing system. The method comprises the following steps: firstly, acquiring three-dimensional point cloud data of a to-be-tested article, and establishing a point cloud model of the to-be-tested article after coordinate transformation; secondly, an outlier detection algorithm is used for detecting outlier data points of the point cloud model of the to-be-tested article, an improved sparrow search algorithm is introduced to find an optimal adaptive threshold value, the outlier data points are removed, and a final point cloud model of the to-be-tested article is obtained; based on the final point cloud model of the to-be-tested article, respectively carrying out an impact resistance test, an insulation performance test and a flame retardant performance test to obtain a performance test result; and finally, analyzing a performance test result to obtain a final performance test result of the to-be-tested article, and completing the performance test. According to the method, modeling analysis is performed on the point cloud data of the to-be-tested article, the purpose of performance testing is achieved, and the method is accurate and objective.

Description

Technical Field

[0001] This invention relates to the technical field of performance testing, specifically to a performance testing method and a safety helmet testing system based on data analysis. Background Technology

[0002] Traditional performance testing methods only perform physical tests on the surface or interior of the item being tested, such as heat resistance tests and puncture tests. These methods cannot adequately consider all the performance characteristics of the item being tested, leading to biases in performance testing. Furthermore, they do not utilize technologies such as artificial intelligence. When testing safety helmets, there are problems such as incomplete testing methods and test errors, which pose safety hazards to production. Summary of the Invention

[0003] In view of the problems in related technologies, the present invention provides a performance testing method and a safety helmet testing system based on data analysis to overcome the above-mentioned technical problems existing in the existing related technologies.

[0004] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0005] This invention relates to a performance testing method based on data analysis, comprising the following steps:

[0006] S1. Obtain the item to be tested, collect the three-dimensional point cloud data of the item to be tested, obtain the point cloud data set of the item to be tested, and establish the point cloud model of the item to be tested after coordinate transformation based on the point cloud data set of the item to be tested.

[0007] S2. Based on the outlier detection algorithm, detect the outlier data points of the point cloud model of the item to be tested, obtain an adaptive threshold, and introduce an improved sparrow search algorithm to find the optimal adaptive threshold of the outlier detection algorithm, thus obtaining an improved outlier detection algorithm. Then remove the outlier data points to obtain the final point cloud model of the item to be tested.

[0008] S3. Based on the final point cloud model of the item to be tested, conduct impact resistance test, insulation performance test and flame retardant performance test respectively, and obtain the impact resistance test results, insulation performance test results and flame retardant performance test results in sequence.

[0009] S4. By comprehensively analyzing the impact resistance test results, insulation performance test results, and flame retardant performance test results, the final performance test results of the item to be tested are obtained, and the performance test is completed.

[0010] This invention generates a surface image of the test item by collecting point cloud data, performs noise reduction processing, and then converts the image coordinate system of the test item to the world coordinate system to generate a point cloud model of the test item. The noise reduction process first filters the point cloud data, reducing the influence of the environment and acquisition device, effectively reducing errors, and modeling the data for easy performance testing, while also offering high applicability and simplicity. Secondly, an improved sparrow search algorithm is used to find the optimal adaptive threshold for outlier detection, resulting in an improved outlier detection algorithm. Outlier data points are then removed to obtain the final point cloud model of the test item. The improved sparrow search algorithm finds the global optimum by simulating sparrow foraging behavior, compared to traditional methods. The algorithm possesses enhanced global search capabilities, and the use of a vector enclosing strategy further improves local search capabilities, accelerates convergence, and reduces the time spent finding the optimal adaptive threshold. A second filtering of the point cloud data, while preserving point cloud features, further reduces errors, laying a solid foundation for subsequent performance testing. The final point cloud model of the test item is then subjected to impact resistance, insulation, and flame retardancy tests to obtain performance test results. This method employs multiple and comprehensive performance tests, resulting in highly reliable and applicable performance test results. The performance testing method is simple, addressing the issue of relying on a single performance testing method. Finally, the performance test results are analyzed to obtain the final performance test results for the test item, completing the performance testing process.

[0011] Preferably, step S1 includes the following steps:

[0012] S11. Obtain the item to be tested. Place the item to be tested on a 3D point cloud data acquisition platform. Use a 3D point cloud data acquisition device to scan the surface of the item to obtain the 3D point cloud data of the item to be tested, generate a surface image of the item to be tested, quantize the pixels of the surface image of the item to be tested, and each pixel represents the 3D point cloud data of the item to be tested. Set a sliding window on the surface image of the item to be tested, select the center pixel of the sliding window, and calculate the average gray value of the pixels in the sliding window, which is recorded as the average gray value of the sliding window. Replace the gray value of the center pixel with the average gray value of the sliding window. Move the sliding window to traverse the surface image of the item to be tested until all pixels on the surface image of the item to be tested have been replaced, and obtain the point cloud data set of the item to be tested.

[0013] S12. Establish an image coordinate system on the surface image of the item to be tested, and perform coordinate transformation based on the point cloud data set of the item to be tested to obtain the world coordinates of the item to be tested. The specific steps are as follows:

[0014] S121. In the image coordinate system, the image coordinates (x1, y1) of the test item are obtained based on the point cloud data set of the test item. The coordinates of the three-dimensional point cloud data acquisition device are set to (x2, y2, z2), and the focal length of the three-dimensional point cloud data acquisition device is a. Then, the relationship between the image coordinates of the test item and the coordinates of the three-dimensional point cloud data acquisition device is as follows:

[0015]

[0016] S122. Define the rotation and translation matrices of the 3D point cloud data acquisition device in the world coordinate system. The relationship between the coordinates of the 3D point cloud data acquisition device and its world coordinates is expressed by the rotation and translation matrices as follows:

[0017]

[0018] Where A′ represents the rotation matrix of the 3D point cloud data acquisition device, A″ represents the translation matrix of the 3D point cloud data acquisition device, and (x3,y3,z3) represents the world coordinates of the 3D point cloud data acquisition device.

[0019] By combining the relationship between the image coordinates of the test item and the coordinates of the 3D point cloud data acquisition device, as well as the relationship between the coordinates of the 3D point cloud data acquisition device and the world coordinates of the 3D point cloud data acquisition device, the world coordinates of the test item are obtained.

[0020] S13. Determine the origin coordinates of the item to be tested in the world coordinate system, project the world coordinates of the item to be tested into three-dimensional space, and generate a point cloud model of the item to be tested.

[0021] This invention generates a surface image of the test object by collecting point cloud data of the test object and performs noise reduction processing to reduce the influence of the environment and the acquisition device, effectively reducing errors. Then, the image coordinate system of the test object is converted into the world coordinate system to generate a point cloud model of the test object. The data modeling process facilitates performance testing and has high applicability and simplicity.

[0022] Preferably, step S2 includes the following steps:

[0023] S21. The 3D point cloud data that makes up the point cloud model of the item to be tested is denoted as the point cloud data set to be detected. Based on the point cloud data set to be detected, an outlier detection algorithm is used to detect outliers in the point cloud model of the item to be tested. The specific steps are as follows:

[0024] S211. Treat the point cloud data in the point cloud data set to be detected as point cloud data points. Select any point cloud data point in the point cloud data set to be detected and denot it as the point cloud data point to be detected. Set the nearest neighbor distance of the point cloud data point to be detected as k. Within the nearest neighbor distance k, calculate the distance between the point cloud data point to be detected and other point cloud data points. Then calculate the average distance and the standard deviation distance. The probability density function calculation formula for the point cloud data point to be detected is as follows:

[0025]

[0026] Where F represents the probability density function of the point cloud data points to be detected, x b α represents the distance between the point cloud data point to be detected and the b-th point cloud data point, β represents the average distance between the point cloud data point to be detected and the point cloud data point, and β represents the standard deviation distance between the point cloud data point to be detected and the point cloud data point.

[0027] S212. Set c to represent a constant, and obtain an adaptive threshold χ = α + c·β based on the average distance and standard deviation distance between the point cloud data points to be detected and the point cloud data points. Detect outliers in the point cloud model of the item to be tested based on the adaptive threshold.

[0028] S22. Using the probability density function of the point cloud data points to be detected as the fitness function, the improved sparrow search algorithm is used to optimize the neighbor distance k and the constant c, resulting in the optimized neighbor distance k and the optimized constant c. The specific steps are as follows:

[0029] S221. Consider the adaptive threshold calculation process as a search space containing a sparrow population. Individual sparrows in the population represent the adaptive threshold. The sparrow population size is m, and the individual sparrow dimensions are n. Initialize the sparrow population and divide it into discoverers, followers, and vigilants. Set the current iteration count to d, and the maximum iteration count to d. At the d-th iteration, the position of the i-th sparrow in the j-th dimension is... The compensation parameter is δ, g1 represents a random number in the interval [0, 1], g2 represents the safety threshold, g3 represents a normally distributed random number, and B represents a matrix of all 1s. The location of the discoverer is updated. When g1 > g2, the position of the i-th sparrow in the sparrow population in the j-th dimension is obtained at the d-th iteration. When g1 ≤ g2, the position of the i-th sparrow in the j-th dimension during the d-th iteration is...

[0030] Replace the follower position update strategy with a centralized strategy, randomly selecting sparrow individual positions from among the discoverers. Sparrow individual location and the location of individual sparrows As the search center, followers move towards it, with a perturbation amplitude of ε. The search center's current position is... Let the follower movement probability be φ, and g4 represent a random number in the interval [0, 1]. When g4 ≤ φ, the position of the i-th sparrow in the sparrow population in the j-th dimension is obtained at the (d+1)-th iteration. Otherwise, the position remains unchanged; the fitness function value corresponding to the current position of the individual sparrow is calculated, and the current best fitness function value is selected. The corresponding position of the individual sparrow is the current best adaptive threshold.

[0031] S222. Set the position of the sparrow corresponding to the current best fitness function value as C. best (d) The position of the sparrow corresponding to the current worst fitness function value is C. worst (d) Based on the individual sparrow position C best (d) and the individual sparrow position C worst (d) Set up a vector enclosing circle in the search space to ensure that the followers update the sparrow individual positions within the vector enclosing circle;

[0032] Update the location of the vigilant, and let g5 represent a random number in the interval [-1, 1]. Let f1 be a constant, and let C be the fitness function value of the i-th sparrow in the j-th dimension at the d-th iteration. worst (d) Corresponding fitness function value f3, sparrow individual position C best (d) The fitness function value f2 corresponding to the sparrow's position is used again. Update the code when f1 > f2. When f1 = f2 At this point, the entire update process of the (d+1)th iteration is completed, and the next generation of sparrow population is generated to continue iterating until the current iteration number reaches the maximum iteration number. Then, the iteration stops, and the final sparrow individual position is obtained. The final sparrow individual position is the optimized neighbor distance k and the optimized constant c.

[0033] S23. Based on the optimized nearest neighbor distance k and the optimized constant c, calculate the optimal adaptive threshold to obtain the improved outlier detection algorithm. Within the nearest neighbor distance k, when the average distance between the point cloud data point to be detected and the point cloud data point is greater than the optimal adaptive threshold, the point cloud data point to be detected is regarded as an outlier data point and removed. Otherwise, the point cloud data point to be detected is a normal data point. This process continues until the three-dimensional point cloud data in the point cloud model of the item to be tested is detected, and the final point cloud model of the item to be tested is obtained.

[0034] This invention uses an improved sparrow search algorithm to find the optimal adaptive threshold for outlier detection, and then removes outlier data points to obtain the final point cloud model of the item to be tested. The improved sparrow search algorithm finds the global optimum by simulating sparrow foraging behavior, and has better global search capabilities than traditional algorithms. The vector enclosing strategy further improves the local search capability, speeds up the convergence, and reduces the time spent finding the optimal adaptive threshold. The point cloud data is then filtered a second time while preserving the point cloud features, further reducing errors and laying a good foundation for subsequent performance testing.

[0035] Preferably, step S3 includes the following steps:

[0036] S31. Select several sample points on the final point cloud model of the item to be tested to obtain a sample point cloud data point set B = {e1, e2, e3, ..., e h}, where e h This represents the h-th sample point cloud data point on the final point cloud model of the item to be tested; the impact resistance performance test is performed based on the set of sample point cloud data points to obtain the impact resistance performance test results. The specific steps are as follows:

[0037] S311. Obtain several test items, select point cloud data points from the sample point cloud data point set to form an impact-resistant point cloud data point set, and ensure that the impact-resistant point cloud data point set is evenly distributed on the top of the test items; mark point cloud data points at different positions on each test item, place the test items sequentially on the test platform, and use the same weight to freely drop and impact the top of the test items, keeping the center of gravity of the weight and the point cloud data points in the impact-resistant point cloud data point set on the same straight line, use a strain acquisition device to record the impact stress wave experienced by the test items, calculate the impact force experienced by the test items based on the impact stress wave, obtain the first impact force set; and record whether there are cracks in the test items, up to the first crack record;

[0038] S312. Obtain several new test items. Select point cloud data points from the sample point cloud data point set and distribute them evenly around the test items. Mark the point cloud data points at different positions on the test items. Place the test items sequentially on the test platform and use a pendulum of the same weight to impact the area around the test items, keeping the impact height the same. Use a strain acquisition device to record the impact stress wave experienced by the test items. Calculate the impact force experienced by the test items based on the impact stress wave to obtain a second impact force set. Record whether there are cracks in the test items to obtain a second crack record. Combine the first impact force set and the second impact force set to obtain the impact force set of the test items. Combine the first crack record and the second crack record to obtain the crack record of the test items. Combine the impact force set and the crack record of the test items to obtain the impact resistance performance test result.

[0039] S32. Obtain several new items to be tested, select point cloud data points from the sample point cloud data point set to form an insulation point cloud data point set, and ensure that the insulation point cloud data point set is evenly distributed on the items to be tested; apply voltage to any point cloud data point in the insulation point cloud data point set, measure the current at other point cloud data points, calculate the resistance corresponding to the point cloud data point, and obtain the resistance of the item to be tested; test the resistance at all point cloud data points in the insulation point cloud data point set in sequence to obtain the resistance set of the item to be tested, and record the resistance set of the item to be tested as the insulation performance test result;

[0040] S33. Obtain several new test items, select point cloud data points from the sample point cloud data point set to form a flame-retardant point cloud data point set, and ensure that the flame-retardant point cloud data point set is evenly distributed on the test items. Place the test items sequentially on the test platform, use a combustion device to vertically burn the test items, record the mass of the test items after combustion, and calculate the combustion loss ratio of the test items based on the mass of the test items before combustion. Calculate the combustion loss ratio corresponding to all point cloud data points in the flame-retardant point cloud data point set sequentially to obtain the combustion loss ratio set of the test items, and record the combustion loss ratio set of the test items as the flame-retardant performance test result.

[0041] This invention conducts impact resistance, insulation, and flame retardancy tests on the point cloud model of the final test item. Through multiple and comprehensive performance tests, the reliability and applicability of the performance test results are high. The performance testing method is simple, solving the problem of a single performance testing method and obtaining accurate performance test results.

[0042] Preferably, step S4 includes the following steps:

[0043] S41. Based on the impact resistance test results, set an impact force threshold and a crack threshold. For the set of impact forces on the test item in the impact resistance test results, if the impact force of the test item in the set of impact forces is greater than or equal to the impact force threshold, the performance test result of the test item is unqualified; if the number of cracks in the crack record of the test item is greater than or equal to the crack threshold, the performance test result of the test item is unqualified; otherwise, the performance test result of the test item is qualified.

[0044] S42. Based on the insulation performance test results, set a resistance threshold. When the resistance of the test item in the insulation performance test results is less than or equal to the resistance threshold, the performance test result of the test item is unqualified. Based on the flame retardant performance test results, set a loss ratio threshold. When the combustion loss ratio of the test item in the flame retardant performance test results is greater than or equal to the loss ratio threshold, the performance test result of the test item is unqualified; otherwise, the performance test result of the test item is qualified. Obtain the final performance test result of the test item and complete the performance test.

[0045] The present invention also discloses a safety helmet testing system, which specifically includes: a point cloud model building module, an outlier data point detection and removal module, a performance testing module, and a performance test result analysis module;

[0046] The point cloud model building module is used to acquire the three-dimensional point cloud data of the safety helmet, and then build the point cloud model of the safety helmet after coordinate transformation.

[0047] The outlier detection and removal module is used to improve the outlier detection algorithm and remove outlier data points after obtaining the optimal adaptive threshold.

[0048] The performance testing module is used to perform impact resistance testing, insulation performance testing, and flame retardancy testing to obtain performance test results.

[0049] The performance test result analysis module is used to analyze the performance test results to obtain the final performance test results of the safety helmet.

[0050] The present invention has the following beneficial effects:

[0051] 1. This invention collects point cloud data of the test item and performs noise reduction processing to reduce the influence of the environment and the acquisition device, effectively reducing errors; then, it converts the image coordinate system of the test item into the world coordinate system to generate a point cloud model of the test item, thus modeling the data for easy performance testing, while also having high applicability and simplicity.

[0052] 2. This invention uses an improved sparrow search algorithm to find the optimal adaptive threshold for outlier detection, and finds the global optimal solution by simulating sparrow foraging behavior. Compared with traditional algorithms, it has better global search capabilities. Furthermore, the use of a vector encirclement strategy improves local search capabilities, accelerates convergence speed, and reduces the time spent finding the optimal adaptive threshold.

[0053] 3. This invention removes outlier data points by using an optimal adaptive threshold to obtain the final point cloud model of the item to be tested. The point cloud data is then filtered a second time while preserving the point cloud characteristics, which further reduces errors and lays a good foundation for subsequent performance testing.

[0054] 4. This invention conducts impact resistance, insulation, and flame retardancy tests on the point cloud model of the final test item. Through multiple and comprehensive performance tests, the reliability and applicability of the performance test results are high. The performance testing method is simple, solving the problem of a single performance testing method and obtaining accurate performance test results.

[0055] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0057] Figure 1 This is a schematic diagram of the performance testing process of the safety helmet testing system provided by the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] In the description of this invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inner", etc., which indicate orientation or positional relationship, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the invention.

[0060] Example 1

[0061] Please refer to Figure 1 The present invention provides a performance testing method based on data analysis, comprising the following steps:

[0062] S1. Obtain the item to be tested, collect the three-dimensional point cloud data of the item to be tested, obtain the point cloud data set of the item to be tested, and establish the point cloud model of the item to be tested after coordinate transformation based on the point cloud data set of the item to be tested.

[0063] S1 includes the following steps:

[0064] S11. Obtain the item to be tested. Place the item to be tested on a 3D point cloud data acquisition platform. Use a 3D point cloud data acquisition device to scan the surface of the item to obtain the 3D point cloud data of the item to be tested, generate a surface image of the item to be tested, quantize the pixels of the surface image of the item to be tested, and each pixel represents the 3D point cloud data of the item to be tested. Set a sliding window on the surface image of the item to be tested, select the center pixel of the sliding window, and calculate the average gray value of the pixels in the sliding window, which is recorded as the average gray value of the sliding window. Replace the gray value of the center pixel with the average gray value of the sliding window. Move the sliding window to traverse the surface image of the item to be tested until all pixels on the surface image of the item to be tested have been replaced, and obtain the point cloud data set of the item to be tested.

[0065] S12. Establish an image coordinate system on the surface image of the item to be tested, and perform coordinate transformation based on the point cloud data set of the item to be tested to obtain the world coordinates of the item to be tested. The specific steps are as follows:

[0066] S121. In the image coordinate system, the image coordinates (x1, y1) of the test item are obtained based on the point cloud data set of the test item. The coordinates of the three-dimensional point cloud data acquisition device are set to (x2, y2, z2), and the focal length of the three-dimensional point cloud data acquisition device is a. Then, the relationship between the image coordinates of the test item and the coordinates of the three-dimensional point cloud data acquisition device is as follows:

[0067]

[0068] S122. Define the rotation and translation matrices of the 3D point cloud data acquisition device in the world coordinate system. The relationship between the coordinates of the 3D point cloud data acquisition device and its world coordinates is expressed by the rotation and translation matrices as follows:

[0069]

[0070] Where A′ represents the rotation matrix of the 3D point cloud data acquisition device, A″ represents the translation matrix of the 3D point cloud data acquisition device, and (x3,y3,z3) represents the world coordinates of the 3D point cloud data acquisition device.

[0071] By combining the relationship between the image coordinates of the test item and the coordinates of the 3D point cloud data acquisition device, as well as the relationship between the coordinates of the 3D point cloud data acquisition device and the world coordinates of the 3D point cloud data acquisition device, the world coordinates of the test item are obtained.

[0072] S13. Determine the origin coordinates of the item to be tested in the world coordinate system, project the world coordinates of the item to be tested into three-dimensional space, and generate a point cloud model of the item to be tested.

[0073] S2. Based on the outlier detection algorithm, detect the outlier data points of the point cloud model of the item to be tested, obtain an adaptive threshold, and introduce an improved sparrow search algorithm to find the optimal adaptive threshold of the outlier detection algorithm, thus obtaining an improved outlier detection algorithm. Then remove the outlier data points to obtain the final point cloud model of the item to be tested.

[0074] S2 includes the following steps:

[0075] S21. The 3D point cloud data that makes up the point cloud model of the item to be tested is denoted as the point cloud data set to be detected. Based on the point cloud data set to be detected, an outlier detection algorithm is used to detect outliers in the point cloud model of the item to be tested. The specific steps are as follows:

[0076] S211. Treat the point cloud data in the point cloud data set to be detected as point cloud data points. Select any point cloud data point in the point cloud data set to be detected and denot it as the point cloud data point to be detected. Set the nearest neighbor distance of the point cloud data point to be detected as k. Within the nearest neighbor distance k, calculate the distance between the point cloud data point to be detected and other point cloud data points. Then calculate the average distance and the standard deviation distance. The probability density function calculation formula for the point cloud data point to be detected is as follows:

[0077]

[0078] Where F represents the probability density function of the point cloud data points to be detected, x b α represents the distance between the point cloud data point to be detected and the b-th point cloud data point, β represents the average distance between the point cloud data point to be detected and the point cloud data point, and β represents the standard deviation distance between the point cloud data point to be detected and the point cloud data point.

[0079] S212. Set c to represent a constant, and obtain an adaptive threshold χ = α + c·β based on the average distance and standard deviation distance between the point cloud data points to be detected and the point cloud data points. Detect outliers in the point cloud model of the item to be tested based on the adaptive threshold.

[0080] S22. Using the probability density function of the point cloud data points to be detected as the fitness function, the improved sparrow search algorithm is used to optimize the neighbor distance k and the constant c, resulting in the optimized neighbor distance k and the optimized constant c. The specific steps are as follows:

[0081] S221. Consider the adaptive threshold calculation process as a search space containing a sparrow population. Individual sparrows in the population represent the adaptive threshold. The sparrow population size is m, and the individual sparrow dimensions are n. Initialize the sparrow population and divide it into discoverers, followers, and vigilants. Set the current iteration count to d, and the maximum iteration count to d. At the d-th iteration, the position of the i-th sparrow in the j-th dimension is... The compensation parameter is δ, g1 represents a random number in the interval [0, 1], g2 represents the safety threshold, g3 represents a normally distributed random number, and B represents a matrix of all 1s. The location of the discoverer is updated. When g1 > g2, the position of the i-th sparrow in the sparrow population in the j-th dimension is obtained at the d-th iteration. When g1 ≤ g2, the position of the i-th sparrow in the j-th dimension during the d-th iteration is...

[0082] Replace the follower position update strategy with a centralized strategy, randomly selecting sparrow individual positions from among the discoverers. Sparrow individual location and the location of individual sparrows As the search center, followers move towards it, with a perturbation amplitude of ε. The search center's current position is... Let the follower movement probability be φ, and g4 represent a random number in the interval [0, 1]. When g4 ≤ φ, the position of the i-th sparrow in the sparrow population in the j-th dimension is obtained at the (d+1)-th iteration. Otherwise, the position remains unchanged; the fitness function value corresponding to the current position of the individual sparrow is calculated, and the current best fitness function value is selected. The corresponding position of the individual sparrow is the current best adaptive threshold.

[0083] S222. Set the position of the sparrow corresponding to the current best fitness function value as C. best (d) The position of the sparrow corresponding to the current worst fitness function value is C. worst (d) Based on the individual sparrow position C best (d) and the individual sparrow position C worst (d) Set up a vector enclosing circle in the search space to ensure that the followers update the sparrow individual positions within the vector enclosing circle;

[0084] Update the location of the vigilant, and let g5 represent a random number in the interval [-1, 1]. Let f1 be a constant, and let C be the fitness function value of the i-th sparrow in the j-th dimension at the d-th iteration. worst (d) Corresponding fitness function value f3, sparrow individual position C best (d) The fitness function value f2 corresponding to the sparrow's position is used again. Update the code when f1 > f2. When f1 = f2 At this point, the entire update process of the (d+1)th iteration is completed, and the next generation of sparrow population is generated to continue iterating until the current iteration number reaches the maximum iteration number. Then, the iteration stops, and the final sparrow individual position is obtained. The final sparrow individual position is the optimized neighbor distance k and the optimized constant c.

[0085] S23. Based on the optimized nearest neighbor distance k and the optimized constant c, calculate the optimal adaptive threshold to obtain the improved outlier detection algorithm. Within the nearest neighbor distance k, when the average distance between the point cloud data point to be detected and the point cloud data point is greater than the optimal adaptive threshold, the point cloud data point to be detected is regarded as an outlier data point and removed. Otherwise, the point cloud data point to be detected is a normal data point. This process continues until the 3D point cloud data in the point cloud model of the item to be tested is detected, and the final point cloud model of the item to be tested is obtained.

[0086] S3. Based on the final point cloud model of the item to be tested, conduct impact resistance test, insulation performance test and flame retardant performance test respectively, and obtain the impact resistance test results, insulation performance test results and flame retardant performance test results in sequence.

[0087] S3 includes the following steps:

[0088] S31. Select several sample points on the final point cloud model of the item to be tested to obtain a sample point cloud data point set B = {e1, e2, e3, ..., e h}, where e h This represents the h-th sample point cloud data point on the final point cloud model of the item to be tested; the impact resistance performance test is performed based on the set of sample point cloud data points to obtain the impact resistance performance test results. The specific steps are as follows:

[0089] S311. Obtain several test items, select point cloud data points from the sample point cloud data point set to form an impact-resistant point cloud data point set, and ensure that the impact-resistant point cloud data point set is evenly distributed on the top of the test items; mark point cloud data points at different positions on each test item, place the test items sequentially on the test platform, and use the same weight to freely drop and impact the top of the test items, keeping the center of gravity of the weight and the point cloud data points in the impact-resistant point cloud data point set on the same straight line, use a strain acquisition device to record the impact stress wave experienced by the test items, calculate the impact force experienced by the test items based on the impact stress wave, obtain the first impact force set; and record whether there are cracks in the test items, up to the first crack record;

[0090] S312. Obtain several new test items. Select point cloud data points from the sample point cloud data point set and distribute them evenly around the test items. Mark the point cloud data points at different positions on the test items. Place the test items sequentially on the test platform and use a pendulum of the same weight to impact the area around the test items, keeping the impact height the same. Use a strain acquisition device to record the impact stress wave experienced by the test items. Calculate the impact force experienced by the test items based on the impact stress wave to obtain a second impact force set. Record whether there are cracks in the test items to obtain a second crack record. Combine the first impact force set and the second impact force set to obtain the impact force set of the test items. Combine the first crack record and the second crack record to obtain the crack record of the test items. Combine the impact force set and the crack record of the test items to obtain the impact resistance performance test result.

[0091] S32. Obtain several new items to be tested, select point cloud data points from the sample point cloud data point set to form an insulation point cloud data point set, and ensure that the insulation point cloud data point set is evenly distributed on the items to be tested; apply voltage to any point cloud data point in the insulation point cloud data point set, measure the current at other point cloud data points, calculate the resistance corresponding to the point cloud data point, and obtain the resistance of the item to be tested; test the resistance at all point cloud data points in the insulation point cloud data point set in sequence to obtain the resistance set of the item to be tested, and record the resistance set of the item to be tested as the insulation performance test result;

[0092] S33. Obtain several new test items, select point cloud data points from the sample point cloud data point set to form a flame-retardant point cloud data point set, and ensure that the flame-retardant point cloud data point set is evenly distributed on the test items. Place the test items sequentially on the test platform, use a combustion device to vertically burn the test items, record the mass of the test items after combustion, and calculate the combustion loss ratio of the test items based on the mass of the test items before combustion. Calculate the combustion loss ratio corresponding to all point cloud data points in the flame-retardant point cloud data point set sequentially to obtain the combustion loss ratio set of the test items, and record the combustion loss ratio set of the test items as the flame-retardant performance test result.

[0093] S4. By comprehensively analyzing the impact resistance test results, insulation performance test results, and flame retardant performance test results, the final performance test results of the item under test are obtained, and the performance test is completed.

[0094] S4 includes the following steps:

[0095] S41. Based on the impact resistance test results, set an impact force threshold and a crack threshold. For the set of impact forces on the test item in the impact resistance test results, if the impact force of the test item in the set of impact forces is greater than or equal to the impact force threshold, the performance test result of the test item is unqualified; if the number of cracks in the crack record of the test item is greater than or equal to the crack threshold, the performance test result of the test item is unqualified; otherwise, the performance test result of the test item is qualified.

[0096] S42. Based on the insulation performance test results, set a resistance threshold. When the resistance of the test item in the insulation performance test results is less than or equal to the resistance threshold, the performance test result of the test item is unqualified. Based on the flame retardant performance test results, set a loss ratio threshold. When the combustion loss ratio of the test item in the flame retardant performance test results is greater than or equal to the loss ratio threshold, the performance test result of the test item is unqualified; otherwise, the performance test result of the test item is qualified. Obtain the final performance test result of the test item and complete the performance test.

[0097] Example 2

[0098] The present invention also discloses a safety helmet testing system, which specifically includes: a point cloud model building module, an outlier data point detection and removal module, a performance testing module, and a performance test result analysis module;

[0099] The point cloud model building module is used to acquire the three-dimensional point cloud data of the safety helmet, and then build the point cloud model of the safety helmet after coordinate transformation.

[0100] The outlier detection and removal module is used to improve the outlier detection algorithm and remove outlier data points after obtaining the optimal adaptive threshold.

[0101] The performance testing module is used to perform impact resistance testing, insulation performance testing, and flame retardancy testing to obtain performance test results.

[0102] The performance test result analysis module is used to analyze the performance test results to obtain the final performance test results of the safety helmet.

[0103] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0104] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A performance testing method based on data analysis, characterized in that, Includes the following steps: S1. Obtain the item to be tested, collect the three-dimensional point cloud data of the item to be tested, obtain the point cloud data set of the item to be tested, and establish the point cloud model of the item to be tested after coordinate transformation based on the point cloud data set of the item to be tested. S2. Detect outlier data points in the point cloud model of the item to be tested based on the outlier detection algorithm, obtain an adaptive threshold, find the optimal adaptive threshold of the outlier detection algorithm to obtain an improved outlier detection algorithm, remove outlier data points, and obtain the final point cloud model of the item to be tested. S3. Based on the final point cloud model of the item to be tested, conduct impact resistance test, insulation performance test and flame retardant performance test respectively, and obtain the impact resistance test results, insulation performance test results and flame retardant performance test results in sequence. S4. By comprehensively analyzing the impact resistance test results, insulation performance test results, and flame retardant performance test results, the final performance test results of the item to be tested are obtained, and the performance test is completed.

2. The performance testing method based on data analysis according to claim 1, characterized in that, S1 includes the following steps: S11. Obtain the item to be tested, collect the three-dimensional point cloud data of the item to be tested, generate a surface image of the item to be tested, and perform filtering processing on the surface image of the item to be tested to obtain a set of point cloud data of the item to be tested. S12. Establish an image coordinate system on the surface image of the item to be tested, and perform coordinate transformation based on the point cloud data set of the item to be tested to obtain the world coordinates of the item to be tested. S13. Determine the origin coordinates of the item to be tested in the world coordinate system, project the world coordinates of the item to be tested into three-dimensional space, and generate a point cloud model of the item to be tested.

3. The performance testing method based on data analysis according to claim 2, characterized in that, S12 includes the following steps: S121. In the image coordinate system, determine the relationship between the image coordinates of the test item and the coordinates of the 3D point cloud data acquisition device. S122. In the world coordinate system, determine the relationship between the coordinates of the 3D point cloud data acquisition device and the world coordinates of the 3D point cloud data acquisition device. Combine the relationship between the image coordinates of the item to be tested and the coordinates of the 3D point cloud data acquisition device to obtain the world coordinates of the item to be tested.

4. The performance testing method based on data analysis according to claim 3, characterized in that, S2 includes the following steps: S21. The three-dimensional point cloud data that make up the point cloud model of the item to be tested is denoted as the point cloud data set to be detected. Based on the point cloud data set to be detected, the outlier points of the point cloud model of the item to be tested are detected using an outlier detection algorithm. S22. Using the probability density function of the point cloud data points to be detected as the fitness function, the improved sparrow search algorithm is used to optimize the neighbor distance k and constant c in the outlier detection algorithm, so as to obtain the optimized neighbor distance k and optimized constant c. S23. Based on the optimized nearest neighbor distance k and the optimized constant c, calculate the optimal adaptive threshold to obtain the improved outlier detection algorithm. Within the nearest neighbor distance k range, remove outlier data points to obtain the final point cloud model of the item to be tested.

5. The performance testing method based on data analysis according to claim 4, characterized in that, S21 includes the following steps: S211. Select the point cloud data points to be detected from the point cloud data set to be detected, set the nearest neighbor distance of the point cloud data points to be detected as k, and calculate the probability density function of the point cloud data points to be detected within the range of the nearest neighbor distance k. S212. Set c to represent a constant. Obtain an adaptive threshold based on the average distance and standard deviation distance between the point cloud data points to be detected and the point cloud data points. Detect outliers in the point cloud model of the item to be tested based on the adaptive threshold.

6. The performance testing method based on data analysis according to claim 5, characterized in that, S22 includes the following steps: S221. Consider the adaptive threshold calculation process as a search space containing a sparrow population. Individual sparrows in the population represent the adaptive threshold. The sparrow population size is m, and the individual sparrow dimensions are n. Initialize the sparrow population and divide it into discoverers, followers, and vigilants. Set the current iteration count to d, and the maximum iteration count to d. At the d-th iteration, the position of the i-th sparrow in the j-th dimension is... The compensation parameter is δ, g1 represents a random number in the interval [0, 1], g2 represents the safety threshold, g3 represents a normally distributed random number, and B represents a matrix of all 1s. The location of the discoverer is updated. When g1 > g2, the position of the i-th sparrow in the sparrow population in the j-th dimension is obtained at the d-th iteration. When g1 ≤ g2, the position of the i-th sparrow in the j-th dimension during the d-th iteration is... Replace the follower position update strategy with a centralized strategy, randomly selecting sparrow individual positions from among the discoverers. Sparrow individual location and the location of individual sparrows As the search center, followers move towards it, with a perturbation amplitude of ε. The search center's current position is... Let the follower movement probability be φ, and g4 represent a random number in the interval [0, 1]. When g4 ≤ φ, the position of the i-th sparrow in the sparrow population in the j-th dimension is obtained at the (d+1)-th iteration. Otherwise, the position remains unchanged; the fitness function value corresponding to the current position of the individual sparrow is calculated, and the current best fitness function value is selected. The corresponding position of the individual sparrow is the current best adaptive threshold. S222. Set the position of the sparrow corresponding to the current best fitness function value as C. best (d) The position of the sparrow corresponding to the current worst fitness function value is C. worst (d) Based on the individual sparrow position C best (d) and the individual sparrow position C worst (d) Set up a vector enclosing circle in the search space to ensure that the followers update the sparrow individual positions within the vector enclosing circle; Calculate the fitness function value f1 of the i-th sparrow in the sparrow population at the position in the j-th dimension during the d-th iteration, and the sparrow's position C. worst (d) Corresponding fitness function value f3, sparrow individual position C best (d) The fitness function value f2 corresponding to the vigilant is compared with the fitness function value f1 and the fitness function value f2 to update the vigilant position. The entire update process of the (d+1)th iteration is completed, and the next generation of sparrow population is generated to continue iterating until the current iteration number reaches the maximum iteration number. Then the iteration stops and the final sparrow individual position is obtained. The final sparrow individual position is the optimized neighbor distance k and the optimized constant c.

7. The performance testing method based on data analysis according to claim 6, characterized in that, S3 includes the following steps: S31. Select several sample points on the final point cloud model of the item to be tested to obtain a set of sample point cloud data points, and then conduct an impact resistance test to obtain the impact resistance test results. S32. Select point cloud data points from the sample point cloud data point set to form an insulation point cloud data point set, and then perform insulation performance testing to obtain the resistance set of the item to be tested. Record the resistance set of the item to be tested as the insulation performance test result. S33. Select point cloud data points from the sample point cloud data point set to form a flame retardant point cloud data point set, and then conduct a flame retardant performance test to obtain the combustion loss ratio set of the test item. Record the combustion loss ratio set of the test item as the flame retardant performance test result.

8. The performance testing method based on data analysis according to claim 7, characterized in that, S31 includes the following steps: S311. Select point cloud data points from the sample point cloud data point set to form an impact-resistant point cloud data point set, and conduct an impact resistance test on the top of the item to be tested to obtain a first impact force set and a first crack record. S312. Conduct an impact resistance test around the object to be tested to obtain a second impact force set and a second crack record; combine the first impact force set and the first crack record to obtain the impact resistance test result.

9. A performance testing method based on data analysis according to claim 8, characterized in that, S4 includes the following steps: S41. Set the impact force threshold and the crack threshold, and compare the impact resistance test results with the impact force threshold and the crack threshold to determine whether the performance test results of the test item are qualified. S42. Set the resistance threshold and the loss ratio threshold, compare the insulation performance test result with the resistance threshold, compare the flame retardant performance test result with the loss ratio threshold, obtain the final performance test result of the item to be tested, and complete the performance test.

10. A helmet testing system implementing any one of claims 1-9, characterized in that, Specifically, it includes: The module includes a point cloud model building module, an outlier data point detection and removal module, a performance testing module, and a performance test result analysis module. The point cloud model building module is used to acquire the three-dimensional point cloud data of the safety helmet, and then build the point cloud model of the safety helmet after coordinate transformation. The outlier detection and removal module is used to improve the outlier detection algorithm and remove outlier data points after obtaining the optimal adaptive threshold. The performance testing module is used to perform impact resistance testing, insulation performance testing, and flame retardancy testing to obtain performance test results. The performance test result analysis module is used to analyze the performance test results to obtain the final performance test results of the safety helmet.

Images