摘要
针对不均匀光照条件下球体识别误检率问题。充分利用球体的颜色和形状信息。对彩色图像经过图像增强预处理后,通过基于HSV彩色空间模型阈值分割的方法,对球体颜色进行识别。将图像转为灰度图像后,通过自适应阈值分割的方法有效地去除了不均匀光照条件对图像分割的影响。在使用Candy边缘检测算法对边缘进行提取后通过基于霍夫变换找圆的方法找出球体的圆。最终经过两个先验信息特征量的交叉匹配,确定目标球体。实验表明该算法有效提升了不均匀光照条件下球体目标识别的准确率。
Aiming at the problem of false detection rate of ball recognition under uneven illumination condition.Make full use of the ball's color and shape information.After color image preprocessing is performed by image enhancement,identify ball color by threshold segmentation based on HSV color space model.After converting the image to a grayscale image,remove the effect of uneven illumination conditions on image segmentation by adaptive threshold segmentation.After extracting the edge using the Candy edge detection algorithm,the circle of the ball is found by finding the circle based on the Hough transform.Finally,the target ball is determined through the cross matching of two prior information feature quantities.Experiments show that this algorithm can effectively improve the accuracy of ball target recognition under the condition of uneven illumination.
作者
陈卫
邓志良
CHEN Wei;DENG Zhi liang(School of Electronic Information,Jiangsu University of Science and technology,Zhenjiang 212003,China)
出处
《电子设计工程》
2019年第21期177-181,共5页
Electronic Design Engineering
关键词
目标识别
不均匀光照
阈值分割
边缘检测
target recognition
uneven illumination
threshold segmentation
edge detection
