Abstract: Aiming at the current problem of difficult calculation of damage volume of engine blade and large error, a damage volume measurement method of compressor blade based on point cloud is proposed. Firstly, a complete point cloud model and a damage point cloud model are acquired by a structured light scanner, and the defective point cloud is obtained by alignment segmentation. Secondly, the defective point cloud is compared and analyzed with the principal component axes after attitude conversion, layered, sliced and projected to obtain the 2D point cloud contour. Finally, a one-way bi-nearest neighbor search algorithm is proposed to extract the contour of the 2D point cloud in an orderly manner, and the area of the projected surfaces is solved using the coordinate resolution method, and the final volume is obtained by accumulating the product of the area of each layer and the slicing interval. The experimental results show that the first principal component axis-direction slice volume proposed in the article is calculated better, and the contour extraction algorithm is more accurate and efficient than the convex packet extraction method, bidirectional nearest-neighbor search and INPS algorithms, and the average relative error is no more than 0.3% compared with the Geomagic software results, which proves the efficiency and effectiveness of the algorithm.