The safety monitoring and identification results of power construction sites are not ideal.Therefore,a research on automatic safety monitoring of power construction sites based on video image recognition is proposed.By combining non-maximum suppression and adaptive threshold generation,a dual threshold segmentation operation was achieved,and the possibility of edge points was marked using a three digit unique heat code.Through multi-dimensional feature matching and difference evaluation,most safety hazards at construction sites have been effectively identified.In the test results,only a few cases did not reach 100.0%,but the effective recognition frequency exceeded 19 times,significantly improving the accuracy of safety monitoring.
关键词
视频图像识别/电力施工现场/非极大值抑制/自适应阈值/双阈值分割操作/多维拟合
Key words
video image recognition/electric power construction site/non-maximum suppression/adaptive threshold/dual threshold segmentation operation/multi-dimensional fitting
维普
万方数据