Repairing Method of Missing Area of Dairy Cows’ Point Cloud Based on Improved Cubic B-spline Curve

Abstract:

The dimensional and size information contained in the 3D point cloud of dairy cows are of great importance to the body evaluation, size measurement and health assessment. While capturing the point cloud of dairy cows from the aisle between the dairy farm and the milking parlor, the missing area on the point cloud happens frequently, which greatly affects the accuracy of the three-dimensional modeling of the dairy cows and the extraction of body size parameters. In order to fix the large missing area on the point cloud, a cubic B-spline curve based method was proposed to repair it. Firstly, the surrounding background was removed from the dairy cow’s point cloud acquired by Kinect v2. Then, the extracted dairy cow’s point clouds were sliced and projected along the x-axis direction of the point cloud coordinate system. In each slice points, some points were filled in the adjacent points with larger spacing. Finally, cubic B-spline curves were used for fitting the filled slice points. The optimal values of parameters h and L were analyzed in the experiment, and a total of 225 frames of 45 Holstein dairy cows’ point clouds were repaired by using the optimal values of parameters h and L. The results showed that the average frame approximation error was reduced by 26.7%, and the large missing area on the point clouds was repaired and the sparseness of point clouds was also improved. The proposed algorithm had better uniformity and approximation performance than the cubic B-spline method, which provided an effective method for repairing large missing point cloud area.