A Method for Identifying Vegetation Under Distribution Power Lines by Remote Sensing

One of the major causes of interruption in distribution power lines is the vegetation encroachment. The vegetation management is challenging and demands efforts in trimming trees planning. The literature presents many methods for encroachment over power lines detection that depends on local installation and manipulation of equipment, which may be unfeasible. Thus, the remote sensing raises as an valuable solution. Therefore, this work proposed a remote sensing based method for identification of probable vegetation encroachment over distribution power lines. Since the free satellite images have low resolution considering the size of treetops, and the high-resolution ones are expensive, our method used the Google Earth images. From that images, texture features and support vector machines were used to identify regions with and without vegetation. The accuracy of the method was of 95% and F1-score above 92% for testing and validation datasets. The method is suitable for real-time application in tree trimming planning, in addition to opening up new possibilities for innovation in vegetation management.