Target point cloud segmentation method applied to turbine blade film hole detection

Abstract

A target point cloud segmentation method applied to turbine blade film hole detection comprises the following steps: acquiring spatial three-dimensional point cloud data of a to-be-detected turbine blade through a five-axis laser measurement platform, and carrying out spatial slicing dimension reduction processing on the blade laser point cloud data in the advancing direction according to a measurement track and laser point cloud distribution characteristics of a laser sensor, so as to obtain a target point cloud data of the to-be-detected turbine blade; calculating a straight line fitting residual by combining the gradient distribution characteristics of the point cloud data of each slice, and screening the point cloud slices of the film hole region; on the basis, carrying out detrending processing on the screened point cloud slices to obtain fitting error class horizontal lines of the slices; a target point set on each slice is obtained through threshold value screening, film hole target area point cloud segmentation is realized, then axis direction and aperture size extraction is carried out on the film hole target area point cloud, quality detection of turbine blade film holes is realized, and thus the point cloud segmentation precision is obviously improved.

Description

technical field [0001] The invention relates to a technology in the field of aero-engine manufacturing, in particular to a target point cloud segmentation method applied to the detection of gas film holes in turbine blades. Background technique [0002] In order to ensure the reliability of aero-engines under high mechanical loads, the axial direction and aperture size of the air film holes in turbine blades have strict precision requirements. The air film hole feature detection algorithm based on laser point cloud can perform high-precision detection of air film holes. Quality Inspection. However, the 3D point cloud of the blade space collected by the line laser sensor, in addition to the point cloud data of the gas film hole in the target area, also includes the point cloud of the blade surface and the noise point cloud, which cannot be directly used for the extraction of the geometric feature of the gas film hole space. Therefore, it is also necessary to use the air film...

AI Technical Summary

Problems solved by technology

The edge detection method that takes the geometric edge of the point cloud as the dividing line mostly regards the direction vector or the sudden change of curvature of the point cloud as the area boundary, and then realizes the purpose of point cloud segmentation in the target area, but the algorithm is greatly affected by the edge detection accuracy. Due to the influence of factors such as point cloud noise, normal vector solution accuracy, and curvature error, under-segmentation or over-segmentation are prone to occur.

In addition, by selecting the region growing method of a specific seed point, according to the corresponding growth strategy to grow and expand to the adjacent points until there are no more points that meet the growth strategy, the purpose of point cloud segmentation in the target region can also be achieved, but due to the initial seed point The distribution position and the formulation of the growth strategy, the segmentation accuracy is difficult to guarantee, and there is no mature seed point selection and growth strategy that can be applied to the segmentation of the gas film hole target point cloud.

Images