Weld joint detection and segmentation method and device based on area array structured light 3D vision

The invention discloses a weld joint detection and segmentation method based on area array structured light 3D vision, and the method comprises the steps: obtaining a corresponding weld joint texture map and point cloud data, extracting a weld joint base metal region contour through point cloud to generate a hypothetical weld joint candidate set, and obtaining a base metal outer contour through the texture map for the point cloud, the method comprises the following steps of: extracting a point cloud weld seam shape main body, extracting a weld seam ROI on the basis, identifying and judging whether a weld seam exists in the weld seam ROI through an SVM (Support Vector Machine) based on an HOG (Histogram of Oriented Gradient) feature vector, fitting the surface of a weld seam joint base metal area corresponding to the weld seam ROI, and segmenting and extracting a corresponding point cloud weld seam shape main body in the ROI based on the fitting surface. The invention further discloses a welding seam detection