| When UUV sails near the surface,it often meet the situation that the UUV may meet with ships and offshore obstacles(e.g.,reef and buoy etc.),especially in the coastal waters,port,fishing areas etc.,Moreover,if UUV’s environmenthas poor visibility,such as fog,evening or night,It is likely to collide with the recovery ship or obstacle on the surface,and causes the damage of equipment.This paper in order to deal with the actual situation.Set the background of UUV’s obstacle target collision avoidance when UUV sails near the surface.To achieve UUV target surface collision location information and access to the design of local collision avoidance algorithm of UUV surface infrared image using dual infrared the visual system.The paper uses infrared image preprocessing,sea antenna detection and extraction,object segmentation and feature extraction and matching.The main works of this paper is:(1)The principle of binocular stereo vision,the common coordinate system,camera imaging model and the calibration principle of binocular camera are introduced in detail.Design the hardware of the UUV binocular vision system(include two infrared cameras,image acquisition card).Based on the existing equipment in the laboratory,the calibration of binocular camera is realized by the combination of Zhang’s calibration and distortion correction.(2)Analyze the characteristics of the target and background in the infrared surface image of the UUV,and handle the infrared image denoising and image enhancement processing.According to the image’s characteristics of narrow range of gray-level and poor contrast,this paper proposes an improved algorithm of merging histogram to enhance the validity of the algorithm.(3)Be aimed at the characteristics(declining sea-sky line,intense wave disturbance)of the surge environment UUV sea surface infrared images,analyze and improve the extraction method of traditional sea-sky-line extraction method.A sea-sky-line extraction method is proposedbased on row-mean value gradient method and linear fitting algorithm,experimental results show that the algorithm of the sea-sky-line location is applicable for the surge environment.In the object segmentation,according to the characteristics of the wave disturbance after removal of low image contrast,improved Otsu threshold segmentation method to achieve target segmentation.(4)According to the binary image of the target image after the target segmentation,using the improved FAST-Harris algorithm,the matching of the feature points of the target image is achieved by using the improved algorithm.The feature vectors are generated by the feature descriptors of the operator,and the corner matching is done based on the Euclidean distance method.The contrast matching algorithm and the improved image matching algorithm in this paper verify the effectiveness of the proposed algorithm.(5)The methods of 3D reconstruction and 3D coordinate solution of the binocular stereo vision model are introduced.A simulation experiment was carried out using UUV surface of binocular vision system is designed in this paper,using the above algorithm for object segmentation,contour feature extraction and image matching of the experimental target,and using the least squares method of three-dimensional coordinates,realizing the target position within the allowable error determination,to prove the feasibility of the proposed algorithm.(6)Analysis of the characteristics of UUV water path planning,introduces the basic principle of artificial potential field method and the advantages and disadvantages,according to the local minima problems of the traditional artificial potential field method was improved,using the improved repulsive function,and solves three typical local minima problem by modifying the direction of repulsion.Object 3D information using the previously obtained,will target the circumcircle and puffing,displayed on the two-dimensional sea surface map.The simulation results show that the improved algorithm is effective. |
PDF Full Text Request