Research on recognition algorithm of LSS based on video in airport clearance area

LSS (Low, Small and Slow) targets have characteristics of low cost, simple operation, easy to carry, low take-off requirements, strong sudden takeoff, and difficult to find and handle. With the development of aviation industry, the cases of LSS flying illegally and used in terrorist attacks are increasing day by day. It will not only cause great negative effects on economy, but also have serious influence on the national security and the normal development of the national economy. In order to eliminate the threat of illegal flight of UAV in the airport clearance area to the safe operation of airport, a LSS recognition algorithm based on the deep learning was proposed in the paper. Compared with other detection types, the video detection was used in the algorithm, which has many advantages such as strong visualization, low cost and fast detection speed. By using the Yolov3 object detection model, one of the most popular deep learning methods, the problem of low accuracy of small objects such as LSS can be solved. The experiments show that the Yolov3 model can improve the accuracy for small objects and can effectively detect the LSS. Thus, the necessary precondition for the subsequent LSS countermeasures can be provided, and the safe operation of the airport can be ensured.