| Popis: |
In order to solve the problems of serious electromagnetic interference, high cost and long detection time of radaroptoelectronic traditional airport runways Foreign Objects Debris Detection(FOD) system, a novel FOD detection algorithm was proposed which based on the improved yolov3. The multi-scale detection and feature extraction network were used to improve the learning ability of object features. Meanwhile, the classical image processing and the deep learning technology was adopt, the algorithm has the ability of target autonomous recognition, besides conventional image restoration. The experimental results show that the algorithm has a strong ability of autonomous recognition and environmental adaptability, the time consumption is better than 0.2s, which can meet the real-time detection of foreign objects debris. It has a guiding significance for the next stage of the engineering of pure-photoelectric foreign object detection system. |
