The Application Research Of Airport Runway Fod Detection Algorithm Based On Difference Image

Foreign objects of airport runway detection system usually refers to some foreign substances may damage the aircraft, short for FOD(foreign objects Debris). The FOD on the runway is a great threat to the aircraft, and a sharp object such as a piece of metal or stone may make aircraft blows up through puncturing the tires. Besides, the debris from tires may damage aircraft engines and other vital components. FOD is not only damage to aircraft but also take away precious lives of people, and even bring huge economic losses to the airport, airline and society.The FOD detection systems developed at home and abroad mainly make use of visible light and radar technology to find, identify, position, record and process the FOD on the airport runway. Normally, the airport runway FOD detection system is mainly composed of three parts: video image processing system, radar system and integrated display system.Video image processing system adopts two main FOD detection algorithms: the static FOD detection algorithm and the difference image FOD detection algorithm. The static FOD detection algorithm needs the priori knowledge of the FOD, and requires a lot of samples for training classifier in the first place. The static FOD detection algorithm is sensitive to the size and shape of the foreign body. But the method doesn’t need image registration between background image and the detected image. After the algorithm detects the FOD, it will identify and classify the FOD according to the characteristics of the target. On the other hand, the difference image FOD detection algorithm needs registration between background image and the detected image, however, the method is very simple and efficient. In addition, it’s not sensitive to the shape and size of the foreign body.Because of the two detection algorithms have their own advantages, so this paper decides to make good use of the two algorithms in the video image processing system mentioned in this paper. This article will focus on FOD detection algorithm based on difference image, and the author’s work includes as following:Firstly, in the second chapter, according to the technology and requirements of FOD detection system of the airport, the author participates and designs the PTZ(Pan/Tilt/Zoom) program and image acquisition program and so on. And then this paper makes use of adaboost classifier method to decide whether the current time is part of the detection period, then it will do corresponding processing according to the detection result.Secondly, the third chapter presents a new image registration method based on mark line, the method is applicable to the registration of the image on the airport runway or the road, the method is simple and efficient. Besides, it is able to meet the real-time feature of the system. In order to exact the mark line from the background, this paper adopts the method of histogram equalization waveform analysis based on histogram equalization. This method can keep the advantage of method of histogram waveform analysis, which is less affected by lighting conditions and also can be used in the situation when the image histogram waveform is irregularly.Thirdly, the threshold used in the fourth chapter for difference image segmentation is extracted from the image gray histogram, and this method can avoid the false alarms effectively with the difference of road and illumination.And last, the fifth chapter is the last module of the airport runway FOD detection algorithm based on difference image. The main content is removing the false alarm objects and extracting true objects. In this paper, the author presents a new method of removing the false alarm objects based on morphological way, which can remove the false alarm objects and retain the true target at the same time. In addition, this chapter also proposes a background image update mechanism, which enables the background image has real-time updates with different lighting conditions of a day.Experiments result shows that the difference image FOD detection algorithm is simple, which has better versatility, stability, and can detect the small target with fast detection speed and high detection accuracy. Furthermore, it can meet the needs of actual runway real-time detection.