Industrial part rapid pose estimation method based on deep learning and point cloud

• Main Authors: LU HUIMIN, ZHENG YUCHAO, CAI JINTONG, LI YUJIE
• Format: Patent
• Language: Chinese; English
• Published: 11.08.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses an industrial part rapid pose estimation method based on deep learning and point cloud, and the method comprises the steps: building a grabbing simulation scene, and collecting a three-dimensional point cloud training data set; constructing an end-to-end pose estimation network based on the point cloud, and training a model by using the three-dimensional point cloud data set to obtain a pre-training model; during online detection, a simulated scene camera is used for shooting a to-be-predicted industrial part, and three-dimensional point cloud representation of the current part is obtained; loading the pre-training model, and inputting the three-dimensional point cloud representation data of the current part into the pose estimation network for calculation; and processing a calculation result of the pose estimation network, and outputting accurate position and pose information of the industrial part. Training data generation, pose estimation task test and evaluation work are completed