摘要
针对足球机器人目标识别的实际特点,以智能足球机器人为研究对象,提出了一种基于足球颜色、面积匹配模板和边缘特征相结合的足球检测方法。此方法首先利用颜色特征信息对图像中颜色相似度进行计算来完成足球的颜色识别。然后,通过分析计算足球在全景摄像头中处于不同位置所成的图像像素尺寸与场地中所处位置距离之间关系,建立了足球识别的面积匹配模板,根据该模板可以计算出在图像中处于不同距离的足球轮廓面积。在此基础上,结合足球边缘特征进一步完成对足球的识别。实验表明该方法能够有效地提高足球识别的效率和准确性,并具有较好的实时性;同时对其他目标物的识别也具有一定的借鉴意义。
According to the actual characteristics of target recognition, intelligent robot is taken as the research object, and an approach to detect football based on the combination of football color, area matching template and edge feature is presented. First, the color recognition of football is com pleted through computing the degree of color similarity of image color characteristics. Then an arez matching template for football detection was established by analysis of the mutual relationship be tween football pixels size and football positions in the image of panoramic camera. The outline area of the football in different distance can be calculated with the template. On the basis above, furthel detection is accomplished by recognition of football edge characters. The experimental results prow that the method can enhance the effectiveness of the football detection and improve the accuracy of football recognition, and has good realtime capability. The method developed could be used for rec oznition of other obiects.
出处
《广西大学学报(自然科学版)》
CAS
CSCD
北大核心
2012年第5期913-920,共8页
Journal of Guangxi University(Natural Science Edition)
基金
教育部中央高校基本科研业务费专项资金资助(CHD2011018)
教育部国家级大学生创新创业训练计划项目(201210710023)
关键词
机器视觉
足球检测
颜色特征
面积匹配模板
边缘检测
machine vision
football detecting
color features
area matching template
edgedetec