Research On Identification Of Micro-UAV Obstacles In Foggy Weather

In recent years,with the rapid development of science and technology,unmanned aerial vehicles(UAV)projects have developed rapidly.The UAV with autonomous flight capability is widely used in military,civil and scientific research because of its advantages of small volume,low cost and flexibility.However,the existing miniaturized UAV can not perceive the information of the surrounding environment independently in the bad environment of fog and low air in the city.The visual range can be controlled by human control,and the obstacle beyond the range of vision will threaten the security of the UAV.Therefore,how to perceive environmental information and ensure the identification accuracy of micro UAV obstacles has become the focus and difficulty of today's society.The main research work in this paper:(1)Image denoising and defog processing.In this paper,the image demogging algorithm based on dark original color prior is used.Then the noise reduction and smoothing the image with the Gauss filter instead of He soft matting algorithm.Finally,the fogging image is obtained with the peak signal noise ratio(peak signal noise ratio),and the fog free image is clearly obtained.(2)The selection and distance measurement of ultrasonic sensors.In this paper,first of all,ultrasonic theory is studied.Then we compare the performance of non-contact sensors and select the appropriate sensors.Then we analyze the principle of ultrasonic ranging and the main factors that affect the ultrasonic ranging,and select the appropriate models for ranging.(3)Optical flow calculation.In this paper,the basic theory of optical flow field is introduced,and the relationship between the light flow field and the motion field is clarified.Then,according to the assumption of constant brightness,continuous time and space consistency,we select the confidence points,then calculate and evaluate the optical flow.Finally,the simulation results are analyzed and compared to verify the effectiveness of the algorithm.(4)Miniaturized unmanned aerial vehicle(UAV)obstacle identification.Pyramid distributed detection is mainly carried out.The data and image de-noising,the size of the optical flow and the set threshold calculated by the ultrasonic sensor are compared to determine whether it is an obstacle,and then the simulation results are analyzed to verify the feasibility of the method.