快速图像分割算法及在MSRS NAO中的应用  

Fast Image Segmentation Algorithm and Its Application in MSRS NAO

作　　者：程建[1] 

机构地区：[1]安徽理工大学计算机学院

出　　处：《安徽理工大学学报（自然科学版）》2013年第1期38-41,共4页Journal of Anhui University of Science and Technology:Natural Science

基　　金：安徽理工大学引进人才基金资助项目

摘　　要：针对MSRS NAO仿真机器人的彩色图像分割问题,提出了一种改进的快速颜色阈值分割算法。首先利用手工颜色标定对被标识物采集其最大、最小值、出现频率最大的HSB值,相对于自动颜色标定的方法更加准确,针对性更强。为了简化计算,进一步提出了元素值为布尔型的索引表,通过逻辑与运算为真来标识物体;最后设计了一个类型矩阵,元素的值不仅可以标识每种物体,而且可以方便地进行逻辑与和逻辑或运算,并给出了针对多个物体的标识算法。实验结果表明,改进后的算法对多个物体的标识,准确性得到了很大提高,改善了算法的时间复杂度,速度提高了近4倍。Aiming at colorful image segmentation problems of MSRS NAO simulation robot, an improved fast colorthreshold segmentation algorithm was proposed. At first, by artificial color calibration maximum value, minimum value, and HSB values of maximum frequency of identified objects were collected. Comparing with automatic col- or calibration, it is more accurate and targeted. In order to simplify calculations, a matrix with Boolean type members was used to identify the object by true result of logical AND operation. Finally, a type matrix was de-signed, elements of which can identify each object and easily make logical AND and logical OR operation; it pro- vides algorithm for multiple objects identification. The experimental results showed that the improved algorithm for multiple object identification greatly improved accuracy and time complexity, and the calculation speed is 4 times faster than by the original algorithm.

关 键 词：MSRS NAO 颜色阈值 手工颜色标定 快速分割算法 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]