基于自适应链码的棉种周长测量  

On detection of cottonseed perimeter based on self-adaptive freeman code

作　　者：刘韶军[1] 王库[1] 

机构地区：[1]中国农业大学信息与电气工程学院,北京100083

出　　处：《计算机应用》2008年第11期2881-2883,共3页journal of Computer Applications

摘　　要：在棉种实时检测系统中,为了识别破损棉种,必须提取棉种的边缘并计算其周长。传统的边缘提取方法有Sobel、Robert等微分算子,以及数学形态学方法和小波变换方法。但是在实时系统中,图像处理速度是影响实时性的关键因素。Freeman链码是图像处理中表示轮廓形状的方法之一,由提取的链码可以得到曲线(包括直线)的曲率、曲线的特征点等,而这些特征是图像压缩、图像重建、图像匹配的依据。由Freeman链码方法提取的图像边缘是单像素的,因此采用这种方法来计算棉种的周长。通过与Sobel、Robert等边缘提取方法进行比较,可以证明运用链码方法求周长,其搜索时间短,计算准确,尤其适用于以快速准确为目标的实时系统。To detect broken cottonseeds in cottonseeds real-time detection system, edge must be extracted and perimeter be calculated. Traditional methods of edge extraction such as Sobel, Robert, mathematical morphology and wavelet transformation are limited in real-time system due to their slow processing rate. Freeman code is one of the methods of edge description. The difference between Freeman code and other methods is that Freeman code gives the curvature and feature points and so on through relative techniques. Meanwhile, the curvature and feature points are essential to image compression and image restoration. The method of Freeman code in edge extraction was applied to cottonseed perimeter calculation. By comparison with Sobel and Robert, Freeman code method is single-pixel and proved to be quick and precise in calculating perimeter, especially in real time system.

关 键 词：FREEMAN链码 边缘提取 图像处理 周长计算 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]