基于YCbCr颜色分割与特征分析的变电站压板状态识别

Substation Platen State Recognition Based on YCbCr Color Segmentation and Feature Analysis

针对传统保护压板图像识别方法对光照条件要求苛刻﹑识别率低以及目前较热的深度学习识别保护压板状态的方法对计算机硬件要求高﹑样本需求量大的问题,提出了一种改进的颜色提取和特征识别压板状态的方法。该方法先对巡检机器人移动端采集到的屏柜图进行图像预处理,提高图像的清晰度和辨识度;然后将彩色图像转换至YCbCr颜色空间中进行颜色特征提取,获取二值图像,对二值图像进行形态学处理并提取连通域;接着对连通域进行形态特征分析,利用面积及边界信息分析除去标识牌和干扰小光斑,利用极限腐蚀算法判别粘连连通域并对粘连连通域进行分割;最后采用简单最小外接矩形和面积最小外接矩形算法获取连通域的外接矩形,对压板进行定位,并根据外接矩形的长高比和偏转角度结合判断压板的状态。试验结果表明所提方法对压板切割的准确性,对复杂环境特殊情况识别的一般性,以及识别方法的高效性和准确性。

Aiming at the problems that the traditional protection platen image recognition method has strict requirements for lighting conditions and low recognition rate,and the current hot deep learning method for identifying the protection platen state has high requirements for computer hardware and large sample size.An improved method of color extraction and feature recognition of platen state is proposed.The screen and cabinet diagram collected by the mobile end of the inspection robot is preprocessed to improve the clarity and recognition of the image.Then the color image is transformed into the YCbCr color space for color feature extraction,the binary image is obtained,the binary image is morphologically processed and the connected domain is extracted.Then,the morphological characteristics of the connected domain are analyzed,the signboard and small interference spot are removed by using the area and boundary information analysis,the adhesive connected domain is identified by using the limit corrosion algorithm,and the adhesive connected domain is segmented.Finally,the simple minimum circumscribed rectangle and the area minimum circumscribed rectangle algorithm are used to obtain the circumscribed rectangle of the connected domain,locate the pressing plate,and judge the state of the pressing plate according to the length height ratio and deflection angle of the circumscribed rectangle. The test results show the accuracy of the proposed method for pressing plate cutting, the generality of recognition for special situations in complex environment, and the efficiency and accuracy of the recognition method.