Endoscopic three-dimensional detection method

Abstract

The invention belongs to the technical field of object surface three-dimensional reconstruction, and provides an endoscopic three-dimensional detection method, which comprises the following steps of: converting a polarization reconstruction dense point cloud and a shadow reconstruction dense point cloud of feature matching points under a pixel coordinate system and a sparse point cloud of stereoscopic vision feature matching points under a world coordinate system to construct a minimum error model; and fusing the three to obtain three-dimensional reconstruction coordinates under a world coordinate system. According to the method, the defect that the complete coordinates of the object cannot be obtained through a traditional three-dimensional endoscopic detection method can be overcome, dense point clouds of all the pixel points are obtained, the obtained reconstruction result is closer to an actual object, meanwhile, a to-be-measured curve in the two-dimensional image is obtained through an edge detection method, the to-be-measured curve is mapped into a world coordinate system through the same method to obtain fused three-dimensional point coordinates, and the length value of the three-dimensional curve is obtained through polynomial fitting and Euclidean distance calculation methods.

Description

technical field [0001] The invention belongs to the technical field of three-dimensional reconstruction of object surfaces, and in particular relates to an endoscopic three-dimensional detection method combining stereo vision, polarization reconstruction and shadow reconstruction. Background technique [0002] Endoscopic inspection technology can transmit the internal conditions through narrow channels without dismantling or destroying the inspected parts. It is more and more used in electronics, shipbuilding industry, machinery manufacturing and other fields. It provides convenience for the detection of internal defects of industrial equipment. With the improvement of modern technology, there are more and more products with complex curved surfaces. Therefore, it is of great significance to study the three-dimensional endoscopic inspection technology that can accurately detect the position and size of target defects. [0003] At present, most of the three-dimensional indust...

AI Technical Summary

Problems solved by technology

Therefore, the current 3D detection usage scenarios have great limitations.

Images