Low Slow Small Aircraft Surveillance System Based on Computer Vision

Recently, low slow small (LSS) aircrafts have become more widely used. However, most LSS aircrafts cannot be detected by air traffic control systems due to their small size, slow speed and low height, which would pose threat to public safety. In this paper, we design and implement an LSS aircraft surveillance system based on computer vision by combing single shot multibox detector (SSD) and robust scale-adaptive mean shift for tracking (ASMS) algorithms, hotspot region generation strategy and matching control. In our system, the SSD detects aircraft target, and then the ASMS tracks the target. The hotspot region generation strategy provides some hotspot areas to improve the object recognition effect of the small-scale target. Next, matching control calculates the fit rotation of the optical turret and the appropriate camera angle to surveil the aircraft of different scales, distances and speeds automatically. Field experiments show that the system can be applied to the LSS aircraft identification and tracking to protect airspace security.