摘要
为实时监控压路机的填压覆盖情况,提出一种基于压路机多角度判别的倾轧轨迹面自动检测算法。提出BOF-MCF(bag-of-feature-matched correlation filter)特征参数,实现压路机角度的自动检测。针对不同角度下的压路机,提出多角度压路机ROI粗定位及其轮廓二级自动分割算法,结合LSD(line segment detector),自动检测压路机碾轮轴长/架长,将帧间图像的碾轮轴长/架长进行叠加计算,得到压路机倾轧轨迹面。实验对国家电网施工现场的视频数据进行测试,其结果表明,该算法对压路机的角度识别正确率达到90%以上,对倾轧轨迹面识别的正确率为83.00%-92.72%。
To realize the real-time monitoring of the filling and pressing coverage,an automatic detection algorithm of road roller’s rolling path under multi-angles was proposed.A BOF-MCF(bag-of-feature-matched correlation filter)feature was proposed,to automatically detect the angles of road roller.The automatic segmentation method of road roller with multi-angles was proposed.The ROI was located,and a finer segmentation algorithm of road roller was presented.Combined with the LSD(line segment detector)algorithm,the roller’s wheel axle/frame was automatically detected.The endpoints of road roller’s wheel axle/frame in each frame were connected to form the rolling path.The experiments were implemented on the video database collected from a construction site of the State Grid.The detection accuracy of road roller’s angles is over 90%.The accuracy of the recognition of rolling path is in the range of 83.00%-92.72%.
作者
范荣全
林明星
刘克亮
张劲
刘俊勇
刘友波
FAN Rong-quan;LIN Ming-xing;LIU Ke-liang;ZHANG Jing;LIU Jun-yong;LIU You-bo(Construction Department,State Grid Sichuan Electric Power Company,Chengdu 610041,China;Tianfu New Area Power Supply Company,State Grid Sichuan Electric Power Company,Chengdu 610041,China;Aba Power Supply Company,State Grid Sichuan Electric Power Company,Aba 623200,China;College of Electrical Engineering,Sichuan University,Chengdu 610065,China)
出处
《计算机工程与设计》
北大核心
2021年第8期2324-2333,共10页
Computer Engineering and Design
基金
国家自然科学基金项目(51977133)
国家电网四川省电力公司科技基金项目(52191918003L)。
