Improvement of workpiece recognition algorithm based on machine vision

Abstract

Machine vision is an important branch of artificial intelligence technology. In the industrial production process with high real-time requirements, machine vision's recognition and sorting efficiency of workpieces determines the efficiency of industrial production. Aiming at the problem of high latency in current machine vision algorithms, a method to reduce the dimensionality of 3D data in the point cloud processing and sampling stages is proposed, which can not only ensure that the characteristic data is not lost, but also can quickly process the 3D data And identify and sort the workpiece