农业温室机器人视觉导航技术的研究

发布时间：2018-05-26 08:36

本文选题：农业机器人 + 温室　；参考：《天津理工大学》2017年硕士论文

【摘要】：随着现代都市型农业和智能制造技术的发展,农业机器人成为农业机械装备中智能控制技术应用的典型代表,其视觉导航系统的研究受到越来越多学者的关注。设施温室具有特殊的内部环境(高温、高湿、有毒等)。与农田相比而言,结构化的温室更适合实施精确农业,能有效减少环境污染,提高工作效率,降低农作人员的工作风险。因此,智能型的温室移动机器人已经掀起各国学者研究的热潮。近年来,由于计算机的不断更新换代与图像处理技术的不断发展,视觉导航成为机器人导航的热门研究领域之一。本文以实现农业温室机器人的自主导航为目的,基于温室农作物成行种植的特点,运用图像处理技术,深入研究其导航路径识别算法。1、对农业机器人进行了概述并介绍其发展现状以及农业温室机器人的应用,阐述了本文的选题背景和组织结构。2、综述了机器人的导航方法、视觉导航的分类以及农业机器人视觉导航的发展,详细阐述了视觉导航的技术体系,包括视觉系统的组成、关键技术和工作过程。3、从三个方面(图像采集、图像预处理和图像分割)对图像处理技术展开研究。针对图像去噪问题,研究了改进的中值滤波和小波去噪两种图像去噪算法,达到了去除混合噪声的目的。结果表明,该算法能获得细节完整的图像且具有良好的实时性。4、研究提出了一种新的分割算法。通过对RGB颜色空间各分量算子重新组合比较,在HSI色彩空间对不同光照条件下各分量的均值和标准差进行比较,确定对图像的(G-B)、(R-B)差值图像和H分量用Ostu法分别进行阈值分割,然后再合并、滤波,将植物从背景中分离出来。所提出的分割方法算法简单,可降低光照不均的影响,能适应复杂的温室环境,并准确识别导航目标。5、研究了改进的“多对一”Hough(霍夫)变换算法,准确地提取植物行列中心线,提高了运算速度,降低了存储空间。根据图像中植物的分布信息和机器人的位姿状态就可以实现机器人的准确导航。对本文所研究的视觉导航算法进行了软件上的实现。
[Abstract]:With the development of modern urban agriculture and intelligent manufacturing technology, agricultural robot has become a typical representative of intelligent control technology in agricultural machinery and equipment. More and more scholars pay attention to the research of vision navigation system. Greenhouse facilities have a special internal environment (high temperature, high humidity, toxic etc. Compared with farmland, structured greenhouse is more suitable for precision agriculture, which can effectively reduce environmental pollution, improve working efficiency and reduce the work risk of agricultural workers. Therefore, intelligent greenhouse mobile robot has aroused the upsurge of scholars all over the world. In recent years, with the development of computer and image processing technology, visual navigation has become one of the hot research fields of robot navigation. In order to realize the autonomous navigation of agricultural greenhouse robot, based on the characteristics of crop planting in greenhouse, image processing technology is used in this paper. In this paper, the navigation path recognition algorithm. 1 is deeply studied. The development of agricultural robot and the application of agricultural greenhouse robot are introduced. The background and organization structure of this paper are described, and the navigation method of the robot is summarized. The classification of visual navigation and the development of vision navigation of agricultural robot are described in detail, including the composition of visual system, key technology and working process. Image preprocessing and image segmentation are studied. Aiming at the problem of image denoising, the improved median filter and wavelet denoising algorithms are studied to remove the mixed noise. The results show that the algorithm can obtain the image with complete details and has good real-time. 4. A new segmentation algorithm is proposed. By comparing each component operator of RGB color space and comparing the mean value and standard deviation of each component under different illumination conditions in HSI color space, the threshold value of the difference image and H component of the image are determined by using Ostu method. Then merge, filter, and separate the plants from the background. The proposed segmentation algorithm is simple, can reduce the influence of uneven illumination, can adapt to complex greenhouse environment, and accurately identify navigation target .5. the improved "many-to-one" Hough-Hough transform algorithm is studied. Accurately extracting the centerline of plant column and column improves the operation speed and reduces the storage space. According to the distribution information of plants in the image and the position and pose state of the robot, the accurate navigation of the robot can be realized. The visual navigation algorithm studied in this paper is implemented in software.
【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【相似文献】