A Novel Approach for Porcupine Crab Identification and Processing Based on Point Cloud Segmentation

Abstract

Despite the increasing application of automated processing equipment in commercial seafood industry, such as large-scale Latin fish and snow crab production lines, manual laboring method dominates in current seafood processing, resulting in low production rate and increased cost. Among various types of seafood crabs, porcupine crabs have shown potential for quality marketable crab meat products. However, their long, sharp spines pose significant challenges for manual laboring and thereby call for robust automated system for processing. In this paper, using 3D point cloud data of the porcupine crab as the input, a novel robot-based approach is proposed to generate the robot trajectory for spine removal. This approach has been validated via a simulation example using ROS (Robot Operating Systems). The proposed method can be introduced into many other manufacturing processing, including polishing, painting, grinding and deburring for work pieces with complex surfaces.