Airport Foreign Object Detection Algorithm Based on Binocular Parallax

The full name of FOD [1] is Foreign Object Debris. This paper combines binocular vision difference to realize Foreign Object Debris detection. The principle of Binocular parallax [2] simulates that the human eyes observe the same position from different observation angles, the spatial information obtained by the left and right eyes is different, and the spatial information is described by detecting this difference. In this paper, the detection of Foreign Object Detection is realized by this principle and binocular camera. Firstly, the appropriate mathematical model is established according to the three-dimensional space information of airport runway plane and binocular camera. Then, the feature points of interest in the right camera are extracted by sift detection algorithm, and then the feature points extracted by the right camera are projected into the image of the left camera according to the mathematical model established above. Finally, The Color difference similarity proposed in this paper is used to filter pixels, which can effectively and quickly detect foreign objects in the region. Experiments show that the algorithm can not only effectively detect the foreign object in the plane, but also eliminate some interference in the plane, such as cracks, and the dirt spots with large difference from the plane.

I. Introduction

With the rapid development of aviation industry, aircraft has become one of the daily transportation tools for the public. Due to the high speed of the aircraft during take-off and landing, any foreign object on the runway may pose a serious threat to the safety of the aircraft [3]–[4]. How to detect and remove the foreign object on the runway in time is an important link to ensure the safety of the aircraft. At present, the most common detection method is the most traditional manual hiking inspection. This method is not only inefficient but also difficult to detect new foreign object in time. Therefore, it is urgent to develop a highly accurate and rapid automatic detection method.