Design And Implementation Of An Obstacle Avoidance System For Small Size Unmanned Helicopters Based On Machine Vision

Helicopters have many advantages compared to the fixed-wing aircraft. They are able to take off and land vertically, and hover in the air. In addition, helicopters can do more things which fixed-wing aircraft cannot, such as vertical flight, lateral flight and rapid change in another direction. Moreover, they are not affected by the limitation of the airports, topography, barriers and artificial gullies, and able to do the low-altitude flight. Due to these advantages, helicopters are widely used in many areas of our society, such as urban traffic control, topographic mapping and field reconnaissance.An obstacle avoidance system is a core system of small sized unmanned helicopter. Usually it uses the machine vision system to captures scenes and models the physical scenario. By this way, the function of dynamic obstacle avoidance can be realized, and a flying path to the target point can be generated.This paper proposes a machine-vision based obstacle avoidance algorithm, and conducts experiments. The result analysis shows that this obstacle avoidance algorithm can successfully model the physical scenario in a certain range, capture the static obstacles, and achieve dynamic obstacle avoidance for small size unmanned helicopters.The main research contents and results of this paper are:1) Research imaging principle of binocular. According to the demands of flying experiment, designed a three-eye camera image acquisition system; solve the problem of image noise effectively;2) Use the three-eye camera image acquisition system to capture physical environment which are in the front of a small size unmanned helicopter, and process image information to model the physical scenario. Calculate the size and spatial location of obstacles, and determine the interference of noise around obstacles to small size unmanned helicopter flight paths.3) According to the model of the physical scenario, give a dynamic obstacle avoidance algorithm. Because it is based on local searches, we do not model the entire physical scenario, and only the front physical scenario of small size unmanned helicopters is calculated .4) Make experiment according to the designed algorithm. Analysis and validation experiment data. The obstacle avoidance algorithm can effectively give the model of obstacles in a less complex physical scenario, successfully implement the dynamic obstacle avoidance for small size unmanned helicopters, and improve the flight safety together with reliability of small size unmanned helicopters.