Research on surface reconstruction algorithm based on upsampling fusion α-shape
Lijuan Qin, Shiyi Wang, Xinyu He · 2025
The living standards and quality of life of the Chinese population have been steadily improving. Farmers increasingly seek non-invasive methods to obtain physical indicators of livestock to avoid stress-induced reactions. Similarly, animal welfare professionals aim to measure the physical characteristics of animals without causing them any disturbance. This study leverages point cloud reconstruction technology combined with machine learning techniques to measure the body features of animals. Using a Kinect depth vision camera, point cloud data of pigs is captured and reconstructed. The processed point cloud data is upsampled, and a surface reconstruction algorithm based on upsampling fusion and α-shape is proposed. The pig's volume is calculated using a convex hull algorithm, while its surface area is determined using PCL (Point Cloud Library) functions.