Research on automatic identification and focusing algorithm based on the day blind ultraviolet unmanned cruise corona detection system

摘要

The electronic leakage on power facility can be positioned by detecting the spectrum on solar blind ultraviolet band, produced by corona radiation. Based on this theory, the corona detector has already been developed home and abroad. In contract with manual corona detector, using corona detector with unmanned aerial vehicles (UAV) can not only detect and track the high-voltage power lines automatically in remote areas, but also reduce the consumptions of manpower. But the way to recognize power facilities under the circumstances of high moving speed and co mplex imaging background is the main difficulty on auto corona detector. So this paper puts forward an algorithm which can recognize and focus power facility fast and precisely. This algorithm firstly realizes auto-focus by using Sobel operator as image quality evaluation function and using improved mountain search algorithm as the control law of motor, then locates the power facilities by applying Hough transfer and region grow split algorithm under the circumstance of complex imaging background. The algorithm was tested on DM6467t embedded platform, and the results shows that this algorithm can recognize target accurately, realize auto-focus rapidly and locate the electronic leakage position.