CN-122023265-A - Curved surface damage detection method, system, device and storage medium

CN122023265ACN 122023265 ACN122023265 ACN 122023265ACN-122023265-A

Abstract

The invention relates to the field of defect detection, and particularly discloses a curved surface damage detection method, a system, a device and a storage medium, wherein the method comprises the steps of obtaining a scanning three-dimensional point cloud of a curved surface to be detected; and establishing a reference fitting curved surface corresponding to the curved surface to be measured based on the scanning three-dimensional point cloud. The method comprises the steps of obtaining normal distance from each data point in a scanning three-dimensional point cloud to a reference fitting curved surface, determining three-dimensional damage characteristics corresponding to each grid region in a two-dimensional texture mapping coordinate system based on distribution conditions of the data points in the two-dimensional texture mapping coordinate system and the normal distance corresponding to each data point, and determining damage to a curved surface to be detected corresponding to the grid region when the three-dimensional damage characteristics meet preset damage judging conditions. The method and the device can realize dynamic damage detection of different curved surface scenes by utilizing the algorithm, and do not need to rely on a neural network model, so that cost saving and efficiency improvement of curved surface damage detection are realized.

Inventors

MA PEIXUN
ZHANG XIAOYANG

Assignees

远景能源有限公司

Dates

Publication Date: 20260512
Application Date: 20251230

Claims (12)

1. The curved surface damage detection method is characterized by comprising the following steps: Acquiring a scanning three-dimensional point cloud of a curved surface to be measured; Establishing a two-dimensional texture mapping coordinate system based on the scanning three-dimensional point cloud, and mapping the scanning three-dimensional point cloud to the two-dimensional texture mapping coordinate system; establishing a reference fitting curved surface corresponding to the curved surface to be measured based on the scanning three-dimensional point cloud, and acquiring the normal distance from each data point in the scanning three-dimensional point cloud to the reference fitting curved surface; Determining three-dimensional damage characteristics corresponding to each grid region in the two-dimensional texture mapping coordinate system based on the distribution condition of the data points in the two-dimensional texture mapping coordinate system and the normal distance corresponding to each data point; and when the three-dimensional damage characteristics meet preset damage judgment conditions, determining that damage exists on the part of the curved surface to be detected corresponding to the grid area, wherein the damage judgment conditions dynamically change according to the curvature distribution condition of the curved surface to be detected.
2. The method for detecting curved surface damage according to claim 1, wherein in determining the three-dimensional damage characteristic corresponding to the grid region, the method comprises: and based on the normal distance corresponding to each data point in the grid area, acquiring the difference value between the maximum value and the minimum value in the normal distances corresponding to all the data points, and taking the difference value as the normal depth range corresponding to the grid area, wherein the three-dimensional damage characteristic comprises the normal depth range.
3. The method for detecting curved surface damage according to claim 1, wherein in determining the three-dimensional damage characteristic corresponding to the grid region, the method comprises: acquiring the three-dimensional space distance between the grid region and the adjacent grid region in the curved surface to be measured and the grid center distance between the grid region and the adjacent grid region; and determining adjacent grid gradients corresponding to the grid region based on the ratio of the three-dimensional space distance to the grid center distance, wherein the three-dimensional damage characteristic comprises the adjacent grid gradients.
4. The method for detecting curved surface damage according to claim 1, wherein in determining the three-dimensional damage characteristic corresponding to the grid region, the method comprises: Determining grid point cloud density corresponding to the grid area based on the distribution condition of the data points in the grid area; And determining a point cloud density variance corresponding to the grid region based on the average point cloud density of the scanned three-dimensional point cloud in the two-dimensional texture mapping coordinate system and the grid point cloud density, wherein the three-dimensional damage feature comprises the point cloud density variance.
5. The method of claim 1, wherein the step of establishing the two-dimensional texture mapping coordinate system comprises: acquiring the local curvature change condition of the curved surface to be detected based on the scanning three-dimensional point cloud; Dynamically adjusting grid dimensions in the two-dimensional texture mapping coordinate system based on the local curvature change condition: when the local curvature of the curved surface to be measured is smaller than a preset threshold value, the grid size of the corresponding area in the two-dimensional texture mapping coordinate system is a preset standard size; when the local curvature of the curved surface to be measured is larger than or equal to a preset threshold value, the grid size of the corresponding area in the two-dimensional texture mapping coordinate system is smaller than the preset standard size.
6. The method of claim 1, wherein in establishing the reference fitted surface and obtaining the normal distance of the data points to the reference fitted surface, comprising: filtering noise points in the scanned three-dimensional point cloud, and fitting the reference fitting curved surface based on a quadric surface model; obtaining projection points of each data point on the reference fitting curved surface; And determining the normal distance corresponding to the data point based on the unit normal vector of the reference fitting curved surface at the projection point position and the distance between the projection point and the data point.
7. The curved surface damage detection method according to claim 1, wherein in determining the damage determination condition, comprising: Acquiring the grid center curvature of each grid region in the two-dimensional texture mapping coordinate system; And determining a damage judgment threshold based on the average value of all the grid center curvatures, wherein the damage judgment threshold is positively correlated with the average value of all the grid center curvatures, and the damage judgment conditions comprise the damage judgment threshold.
8. The method of claim 7, wherein the three-dimensional lesion characterization includes at least one lesion characterization parameter; And judging that the three-dimensional damage characteristic accords with the damage judging condition under the condition that the value of any damage characteristic parameter is larger than the damage judging threshold value.
9. The method for detecting curved surface damage according to claim 8, wherein the damage characteristic parameters include a normal depth range, an adjacent grid gradient and a point cloud density variance corresponding to the grid region; the normal depth range is used for representing extreme value difference conditions of normal distances from the data points in the grid area to the reference fitting curved surface; The adjacent grid gradients are used for representing gradient change conditions between the grid areas and the adjacent grid areas; the point cloud density variance is used to characterize a difference between a distribution density of the data points within the grid region and an average distribution density of the scanned three-dimensional point cloud in the two-dimensional texture mapping coordinate system.
10. A curved surface damage detection system, comprising: the scanning acquisition unit is used for acquiring a scanning three-dimensional point cloud of the curved surface to be detected; the coordinate establishing unit is used for establishing a two-dimensional texture mapping coordinate system based on the scanning three-dimensional point cloud and mapping the scanning three-dimensional point cloud to the two-dimensional texture mapping coordinate system; the standard curve fitting unit is used for establishing a standard fitting curve corresponding to the curve to be measured based on the scanning three-dimensional point cloud, and obtaining the normal distance from each data point in the scanning three-dimensional point cloud to the standard fitting curve; The three-dimensional feature determining unit is used for determining three-dimensional damage features corresponding to each grid region based on the distribution condition of data points in each grid region in the two-dimensional texture mapping coordinate system and the normal distance corresponding to each data point; The damage evaluation unit is used for determining that damage exists on the part of the curved surface to be detected corresponding to the grid area when the three-dimensional damage characteristic accords with a preset damage judgment condition, wherein the damage judgment condition is dynamically changed according to the curvature distribution condition of the curved surface to be detected.
11. A curved surface damage detection device, comprising: At least one processor, and a memory communicatively coupled to the at least one processor; Wherein the memory stores instructions executable by at least one of the processors to enable the at least one processor to implement the steps of the curved surface damage detection method of any one of claims 1 to 9.
12. A computer readable storage medium having stored therein computer instructions which when executed by a processor perform the steps of the curved surface damage detection method of any of claims 1 to 9.

Description

Curved surface damage detection method, system, device and storage medium Technical Field The disclosure relates to the technical field of defect detection, and in particular relates to a curved surface damage detection method, a curved surface damage detection system, a curved surface damage detection device and a curved surface damage detection storage medium. Background The curved surface damage detection refers to identifying and analyzing damage or defects on a three-dimensional curved surface through various methods and technologies, and is generally applied to health monitoring links of materials, structures or equipment, for example, the curved surface damage detection can be used for realizing pit detection of a worm wheel blade of a fan, appearance detection of a shell on the surface of an energy storage battery assembly, and the like, and is used for ensuring the safety and performance of the structure. The existing curved surface damage detection generally depends on a special deep learning model, a great deal of time, labor and other costs are required to carry out the marking of training data and the model training process, and meanwhile, in the specific curved surface damage detection process, the carrying and iterative optimization of the model highly depend on the calculation force of a graphic processing unit (Graphics Processing Unit, GPU), so that the curved surface damage detection needs to consume higher hardware cost. Disclosure of Invention The embodiment of the disclosure aims to provide a curved surface damage detection method, a system, a device and a storage medium, which can realize dynamic damage detection of different curved surface scenes by utilizing an algorithm without depending on a neural network model, save the cost of model training, data labeling, hardware configuration and the like, and improve the efficiency of curved surface damage detection. The method for detecting the curved surface damage comprises the steps of obtaining a scanning three-dimensional point cloud of a curved surface to be detected, establishing a two-dimensional texture mapping coordinate system based on the scanning three-dimensional point cloud, mapping the scanning three-dimensional point cloud to the two-dimensional texture mapping coordinate system, establishing a reference fitting curved surface corresponding to the curved surface to be detected based on the scanning three-dimensional point cloud, obtaining a normal distance from each data point in the scanning three-dimensional point cloud to the reference fitting curved surface, determining three-dimensional damage characteristics corresponding to each grid area in the two-dimensional texture mapping coordinate system based on distribution conditions of the data points in the two-dimensional texture mapping coordinate system and the normal distance corresponding to each data point, and determining that damage exists in a curved surface part to be detected corresponding to the grid area when the three-dimensional damage characteristics meet preset damage judging conditions, wherein the damage judging conditions dynamically change according to curvature distribution conditions of the curved surface to be detected. In one possible implementation manner of the first aspect, in determining the three-dimensional damage feature corresponding to the grid area, the method includes acquiring a difference value between a maximum value and a minimum value in normal distances corresponding to all data points based on normal distances corresponding to each data point in the grid area, and taking the difference value as a normal depth range corresponding to the grid area, where the three-dimensional damage feature includes the normal depth range. In one possible implementation manner of the first aspect, in determining the three-dimensional damage characteristic corresponding to the grid region, the method includes obtaining three-dimensional space distances between the grid region and the adjacent grid region in the curved surface to be measured and grid center distances between the grid region and the adjacent grid region, and determining adjacent grid gradients corresponding to the grid region based on a ratio of the three-dimensional space distances to the grid center distances, wherein the three-dimensional damage characteristic includes the adjacent grid gradients. In one possible implementation manner of the first aspect, in determining the three-dimensional damage feature corresponding to the grid area, determining grid point cloud density corresponding to the grid area based on a distribution condition of data points in the grid area, and determining a point cloud density variance corresponding to the grid area based on an average point cloud density and a grid point cloud density of a scanned three-dimensional point cloud in a two-dimensional texture mapping coordinate system, wherein the three-dimensional damage feature comprises the point cl