In the paper an algorithm of determining the gaps in the point clouds obtained after scanning of the processed workpieces by means of vision systems is proposed. The resulting point cloud is projected on a plane, the location of which is set in advance for each type of workpieces. Next, the resulting area of projection is estimated, visible points are allocated and their density on the obtained part of model is determined. In this analysis of the quality of point clouds are proposed to use of relatively simple, well-known and proven algorithms for processing of raster images. The efficiency and effectiveness of this algorithm is confirmed by a model example.