Multi-type welding groove feature extraction method based on three-dimensional point cloud

The invention relates to the technical field of welding groove detection, in particular to a multi-type welding groove feature extraction method based on three-dimensional point cloud, which comprises the following steps: 1, measuring three-dimensional point cloud data of a workpiece; 2, filtering the three-dimensional point cloud data; 3, a cluster set is constructed and classified, and a clustering result is obtained; 4, constructing a piecewise linear regression error function err (r) of each cluster; 5, minimizing the error function err (r) to search the optimal r; 6, comparing the information criterion value under the current segment number with the previous value, judging whether the segment number needs to be updated or not, if so, storing the current optimal r, and circulating the steps 4 to 6; otherwise, returning the optimal r to obtain a fitting curve; and 7, calculating groove feature points. According to the method, the quality of the model is evaluated by using AIC and BIC methods, a multi-type