Recently, intelligent unmanned aerial vehicles (UAVs) have shown great advantages of flexibility and productivity in power line inspection, wherein robust detection of power lines from aerial images for automatic power line following navigation is required. However, identifying power lines accurately from a cluttered background is challenging due to the limited resolution of onboard cameras and the noisy environment. In this letter, we propose a novel power line detection method, denoted by CODNet, for the application of UAV navigation. Unlike existing works, the proposed method can extract features of power lines from cluttered backgrounds automatically and predict centers and orientations of power lines in the scene simultaneously. Besides, we introduce a new clustering method to summarize the average location and orientation of detected power lines as a guide for the automatic navigation of UAVs. Finally, experimental results demonstrate both the effectiveness and the superiority of the CODNet.
The use of object detection methods is becoming increasingly important in the internet of smart energy. Accurate detection of power line region and suspicious objects can reduce the probability of accidents and save manpower. However, the noisy outdoor background and time-varying lighting conditions bring difficulties to those tasks. To address those challenges, this paper investigates mainstream methods of power line region surveillance and proposes a new method for improving the accuracy of power line region and risk object detection by modeling the heat map branch of CenterNet, which is the most representative of anchor-free detectors, with a cascade guiding structure and an improved loss function. Finally, comprehensive experiments are conducted to compare some state-of-the-art detectors and the proposed model on power line surveillance tasks. The test results demonstrate both the effectiveness and superiority of the proposed method.
1 January 2020 in Intelligent Automation and Soft Computing
The rapid development of Internet of Things (IoT) technologies has boosted smart energy networks in recent years. However, power line surveillance systems still suffer from the low accuracy and efficiency of the power line area recognition and risk objects detection. This paper proposes a new customized loss function to tackle the disequilibrium of the size of objects on multi-scale feature maps in the deep learning-based detectors. To validate the new concept and improve the efficiency, we also presented a new object detection model. Experimental results are provided to exhibit the advantage of our proposed method in both accuracy and efficiency.