摘要
视觉系统是仿人对弈机器人研究的关键部分。而视觉系统的难点是关于棋子的边缘检测,本文采用的是Marr算子的小波变换的边缘检测,这种方法用高斯函数的二阶导数作为小波,这是一种具有多尺度的算法,即通过改变参数可实现图像不同范围的平滑处理。实验表明该方法不仅能够依据边缘的奇异度区分不同的边缘,而且很好的抑制了图像噪声,符合人机对弈的机器视觉分级别识别边缘检测的要求。
Vision system of Chinese chess playing robot is the key part. While the difficult point of the vision system is edge detect of chessman, this text apply muhiscales edge detection based on the Marr Wavelet Transform, this method apply the Second derivative of gauss function to wavelet, and then, change a parameter to achieve the purpose of smoothing different range of image. Experiments show that this method cans not only distinct different edge according to the Singular degrees of edge, but also suppress the image noise, meet the requirements of Human-machine game with the visual identification of the sub-level edge detection.
出处
《微计算机信息》
2009年第27期37-38,55,共3页
Control & Automation
关键词
边缘检测
多尺度
小波变换
高斯函数
Edge Detection
Multiscales
Wavelet Transform
gauss function
