Research On System Of Vision-Based For Quadrotor Landing Onto Specific Platform

With understanding of the computer vision system, and the development of various complex algorithms, microprocessor technology and ultra large scale integrated circuit technology, it is possible to calculate the visual application to the four quadrotor. And vision based quadrotor landing has become a hot spot in the specific landing platform. Compared with the traditional method of landing, the landing is more accurate.First of all, this paper presents a visual autonomous landing scheme for quadrotor landing in a specific platform, design a new visual landing logo, detailing software and hardware platform of the visual landing system and algorithms and so on.Then, the detection and tracking of a specific landing platform, which is based on the call of the camera to obtain images, the image is gray and Wiener denoising processing, on the basis of the improved Wellner adaptive threshold algorithm to obtain binary images. Then it is the outline of the search, filter processing, the filtering of the contour of the polygon approximation, using the approximation method is the minimum perimeter polygon(MPP), polygon approximation can be obtained after the top of the outline of the polygon. Secondly, the contour of the polygon is detected by the characteristic of the visual landing. Finally, it is based on the AR model to predict the specific platform.Secondly, the position parameters of the quadrotor are calculated, and the relative displacement of the quadrotor and the specific landing platform is based on the displacement of the image coordinate system. Firstly, the camera is calibrated and the camera’s internal parameters are obtained, and then the pixel displacement can be converted to the actual displacement by using the similar triangle according to the image model. Other flight parameters such as height, yaw angle, roll angle and pitch angle of the four rotor blades can be obtained by airborne ultrasonic wave and inertial measurement unit(IMU).The experiment includes two parts, the first part is based on the vision of the quadrotor landing system on a specific platform. The second part is in the process of implementation the actual acquisition of images in the visual landing algorithm in various steps of the process. Experimental results show that the system can make the quadrotor safely and accurately landing on a specific landing platform.