基于biSCAN和SVM的机器人目标识别新算法研究

Research on a New Algorithm for Robots' Recognition of Objects Based on biSCAN and SVM

针对机器人足球比赛环境容易受光照变化影响,造成目标识别率低、光照自适应性差的问题,采用c1c2c3彩色不变的特征,利用biSCAN算法设置一系列从圆形图像中心出发且与图像半径为同方向的径向扫描线来提取目标所在区域,并用支持向量机(support vector machine,SVM)进行目标识别.实验表明,该方法提高了目标识别率,具有较强的光照自适应能力.

Robot soccer environment is vulnerable to changes in illumination, which results in the low recognition rate, the poor illumination adaptability. In order to solve these problems, it adopts clc2c3 col- or invariant features, uses biSCAN to set a few ways of radial scanning which start from the image center and along the radius to extract the object area, and then uses SVM to recognize the object. Experimental results show that the method improves the object recognition rate and the ability to adapt to illumination.