Design And Research On Vision System For The Airport Runway Detection Robot

Since the airport runway is an very important platform for the aircrafts to take off and land, its safety problem is getting more and more attention with the development of aviation industry. FOD (Foreign Object Debris) and cracks in the airport runway are the biggest threat to the aircraft take-off and land. Therefore, in order to ensure the safety of flights, the airport runway needs regular detection of FOD and cracks. Foreign countries have been studying the automatic detecting equipments of FOD and cracks for many years and as a result, many kinds of real-time monitoring systems are widely used in the airports. However,the domestic research starting later, application of FOD detection system in domestic airports is rarely reported. Therefore, the automatic detection of FOD and cracks in the airport runway equipment has become an urgent need for the domestic airport.Based on the researches of FOD detection system and the development of modern intelligent service robot system, this paper designs a robot vision system of FOD and cracks detection for the airport runway. In the sunny weather daytime, robot vision system uses optical camera to capture runway image, and discriminate the target when vision system detects the target of FOD or cracks.This paper designs hardware platform and actualizes the algorithm that uses on vision system. The hardware platform of vision system uses DSP+FPGA image processing structure which not only gives full play to the advantages of these two processors, but also meets the real-time requirements of image processing. The MATLAB simulation of the whole software algorithm is carried out to verify the validity and performance of the algorithm.When the airport runway images containing marking are detected, the traditional detection method may identify the marking as a goal mistakenly, resulting in false detection.So, this paper designs two improved methods: background subtraction method based on filling and improved edge detection algorithm. In order to accurately verify the detection rate of two algorithms, the paper has carried on the simulation research of the two algorithms. The simulation results show that the detection rate of two kinds of improved algorithm are higher than the traditional target detection algorithm.For the recognition of FOD and cracks in the airport runway, this paper adopts the method of object recognition based feature. In this paper, target recognition classifier based on BP neural network is designed and simulated. Compared with SVM target classifier, the target recognition classifier based on BP neural network has a high recognition rate.At last, the paper carries out the hardware implementation of the algorithm, and completes the hardware debugging. In this paper, the vision system is validated by the experiment. The results show that the vision system can effectively detect and recognize the FOD and the crack in the airport runway under the given experimental conditions. In the hardware environment, the algorithm is debugged. The research of this paper provides a better solution for the foreign matter and crack detection of the airport runway.