低光照场景下10 kV配网带电作业机器人目标识别  

Target recognition of live working robots in 10 kV distribution networks under low light conditions

作　　者：何成 李根 景磊 张杨 刘岳 马亚龙 HE Cheng;LI Gen;JING Lie;ZHANG Yang;LIU Yue;MA Yalong(Ordos Power Supply Branch of Inner Mongolia Electric Power(Group)Co.,Ltd.,Ordos 017010,China)

机构地区：[1]内蒙古电力(集团)有限责任公司鄂尔多斯供电分公司,内蒙古鄂尔多斯017010

出　　处：《电子设计工程》2025年第6期16-20,共5页Electronic Design Engineering

基　　金：国家重点社科基金项目(NMDL20230044)。

摘　　要：在低光照条件下,场景中的对比度降低,导致10 kV配网带电作业机器人无法准确地区分目标和其周围的环境。因此,提出低光照场景下10 kV配网带电作业机器人目标识别方法。计算灰度图像中灰度级出现的概率,分析原图像灰度与新图像灰度关系。设计时间窗口函数,扩展像素强度分布增强图像对比度,利用相似三角形原理,建立笛卡尔坐标内点云分布矩阵。计算点云直线度和主方向与平面法向夹角,构建目标识别模型。由实验结果可知,该方法最高灰度值为195,说明像素亮度高,识别的目标图像清晰。Under low lighting conditions,the contrast in the scene decreases,resulting in the inability of the 10 kV distribution network live working robot to accurately distinguish between the target and its surrounding environment.Therefore,a target recognition method for live working robots in 10 kV distribution networks under low light conditions is proposed.Calculate the probability of grayscale levels appearing in grayscale images and analyze the relationship between the original image’s grayscale and the new image’s grayscale.Design a time window function to expand pixel intensity distribution and enhance image contrast,and use the principle of similar triangles to establish a point cloud distribution matrix in Cartesian coordinates.Calculate the straightness of the point cloud and the angle between the principal direction and the plane normal,and construct a target recognition model.The experimental results show that the highest grayscale value of this method is 195,indicating high pixel brightness and clear recognition of the target image.

关 键 词：低光照场景 10 kV配网 带电作业机器人 目标识别 

分 类 号：TN202[电子电信—物理电子学]