基于机器视觉的烹饪机器人锅具识别算法

全面介绍了基于机器视觉的烹饪机器人锅具识别算法,用于提取复杂背景中的锅具曲线.算法分为两个阶段:特征点确定、图像匹配.使用曲率作为特征点,给出了数字图像的离散特征点计算方法和加权Hausdorff距离识别锅具轮廓特征点.解决了图像因为旋转、位移、缩放造成图像形变以及由于锅具动作图像部分遮挡、动态模糊造成识别困难问题.实验表明该锅具识别算法能够高效地提取画面中的锅具轮廓,能较好地满足烹饪机器人示教系统的要求.

Identification Algorithm for Cooking Robot Wok Based on Machine Vision

A comprehensive introduction to the identification algorithm for cooking robot wok based on machine vision, it is utilized to extract the curve of wok in the image with complex background. The vision identification is divided into 2 stages: characteristic points determination and image matching. The curvature of the edge is used as characteristic points, a new algorithm is used to obtain the discrete characteristic points of digital image, and weighted Hausdorff distance is used to identify the characteristic point of wok edge. The rotation, translation and scale of image will bring about the problem of image deformation and the motion of wok will bring about the problem of partial occlusion and dynamic fuzzy image, the identification algorithm in this paper can solves the these problems. Great deals of experiments show that this method can extract the wok outline from the image availably and satisfy the requirement of studying system for cooking robot.