全自动水草清理作业船视觉辅助导航技术研究

发布时间：2018-06-06 12:27

本文选题：水草清理船 + 机器视觉　；参考：《江苏大学》2017年硕士论文

【摘要】：我国是河蟹等水产品消费的大国,具有巨大的养殖面积和消费市场,但是与发达国家相比,我国河蟹养殖行业存在生产方式落后、成本高、规模小和工业化水平较低的问题。河蟹养殖的水草清理和维护工作还都是依靠人工作业完成,有必要研发出一种全自动化的水草清理船来降低农户的劳动强度和养殖成本。本文设计了一个基于GPS定位系统的视觉辅助导航系统,旨在实现全自动作业船的自动导航功能。本文主要研究工作如下:一、使用明轮动力平台作为实验平台,搭建基于GPS定位的视觉辅助导航系统,并介绍视觉辅助导航系统的设计思路和硬件结构。二、采用数字图像处理技术对采集到的水面水草图像进行分割。首先进行图像的预处理操作,包括降噪滤波和灰度化处理,为后续基于阈值的图像分割处理奠定基础,然后完成对目标(水草)和背景(水面)的分割工作,通过形态学运算和有效区域保留的方法得到导航有效区域。三、根据得到的导航有效区域提取出一条导航基准线。首先比较分析最小二乘直线拟合和霍夫变换两种直线检测法的工作效果,然后根据本课题的应用背景提出一种改进的直线检测算法提取出导航基准线,作为作业船的导航路径。四、采用张氏标定法对CCD工业相机进行标定,得到其内外参数,然后根据摄像机成像模型与相机的内外参数将图像坐标系中的导航基准线转换到世界坐标系中,并计算出航向角。五、使用卡尔曼滤波算法将视觉导航模块和GPS定位模块的导航参数进行信息融合,建立组合导航模型,以提高自动导航系统的定位精度。通过导航实验可以得到,信息融合后实验数据的横向极差、平均偏差和标准误差分别下降了42%、65%和67%。
[Abstract]:China is a big country in the consumption of aquatic products such as crabs, and has a huge breeding area and consumption market. However, compared with the developed countries, the production mode is backward, the cost is high, the scale is small and the industrialization level is low in our country. In order to reduce the labor intensity and the cost of farming, it is necessary to develop a kind of fully automatic cleaning boat of water and grass to reduce the labor intensity and the cost of farming. A visual aided navigation system based on GPS positioning system is designed in this paper. The main research work of this paper is as follows: firstly, the visual navigation system based on GPS positioning is built using the power platform of the open wheel as the experimental platform, and the design idea and hardware structure of the visual aided navigation system are introduced. Secondly, the digital image processing technology is used to segment the collected surface water grass image. First, the image preprocessing operations, including noise reduction filtering and grayscale processing, are carried out to lay the foundation for the subsequent threshold-based image segmentation, and then the target (grass) and background (water surface) segmentation is completed. The navigation effective region is obtained by morphological operation and effective region reservation. Thirdly, a navigation datum is extracted according to the effective navigation area. This paper first compares and analyzes the results of the two methods of least-square line fitting and Hough transform, and then proposes an improved line detection algorithm to extract the navigation datum, which can be used as the navigation path of the operation ship according to the application background of this subject. Fourthly, the CCD industrial camera is calibrated by Zhang's calibration method, and its internal and external parameters are obtained, then the navigation datum in the image coordinate system is converted to the world coordinate system according to the camera imaging model and the camera's internal and external parameters. The heading angle is calculated. Fifthly, the navigation parameters of visual navigation module and GPS positioning module are fused by Kalman filter algorithm, and the integrated navigation model is established to improve the positioning accuracy of automatic navigation system. The results of navigation experiments show that the transverse range, average deviation and standard error of the experimental data after information fusion are decreased by 42% and 67%, respectively.
【学位授予单位】：江苏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S969;TP391.41

【参考文献】