Beef cattle weight prediction method based on point cloud deep learning

The patent name of the invention is a beef cattle weight prediction method based on point cloud deep learning. Beef cattle point cloud data are obtained through a 3D laser camera, and the beef cattle point cloud data are preprocessed by applying fusion filtering, clustering segmentation and curved surface reconstruction. And identifying three feature parts of the cattle by using a beef cattle data set PointNet + + model. And then calculating the three-dimensional coordinates of the geometric center point of the feature part, and calculating the oblique length and the chest circumference by using the coordinates of a spatial Euclidean distance formula. And finally, on the basis of the Johnson method, the body weight of the beef cattle is predicted in combination with a deep learning method, and non-contact efficient measurement of the body size and the body weight of the beef cattle is realized. 该发明专利名为"一种基于点云深度学习的肉牛体重预测方法"。通过3D激光相机获取肉牛点云数据，运用融合滤波、聚类分割和曲面重构对肉牛点云数据进行预处理。利用肉牛数据集PointNet++模型识别出牛的3个特征部位。然后计算特征部位的几