玉米苗基部图像定位方法设计

发布时间：2017-12-31 11:44

本文关键词：玉米苗基部图像定位方法设计　出处：《沈阳农业大学》2017年硕士论文　论文类型：学位论文

【摘要】：传统的人工作业对玉米苗基部进行施肥、注水、撒药在现阶段已经不适用,这样会浪费大量人力和时间,效率很低,工作量极大。因此,利用图像处理的方法对玉米苗基部进行定位,实现田间智能化作业,智能的为玉米苗基部进行施肥、注水、撒药已成为未来的一种趋势。本文在阅读了大量的国内外相关文献的基础上,研究了对玉米苗基部图像定位方法设计的方法和具体操作流程,为今后的智能化田间作业提供了可行性。本文着重从以下几方面进行了研究:(1)对影响图像采集质量的因素和图像的去噪及灰度化处理进行了试验研究。通过图像采集系统对采集高度、角度和光照的研究,试验研究出了较清晰的静态采集图像的三要素,在采集到清晰图像的基础上选择用3×3的均值滤波法对玉米苗图像进行去噪处理;为了保证后续处理的速度尽可能快速、准确,最终选用加权平均值法对玉米苗图像进行灰度化处理。(2)在灰度图像的基础上对图像分割和连通区域提取主干、形态学处理进行了研究。为了使目标信息与背景信息区分更加明显,选择Ostu阈值方法对图像进行分割;为了使目标信息更加简洁,方便后续处理,本文对分割后的二值图像进行连通区域提取主干和形态学处理。(3)对图像的边缘检测方法进行了研究。为了使目标信息清晰完整,本文采用Canny算子对图像进行边缘检测,将边缘图像与二值图像重合再次进行形态学处理,形成最终清晰、完整二值图。(4)对玉米苗基部定位方法和生长角、茎粗和茎杆识别方法进行研究。通过几种定位方法对比研究,选择旋转投影定位法对玉米苗基部进行定位处理,在图像旋转的基础上提取出了生长角,并对茎粗和茎杆进行了识别。
[Abstract]:Traditional artificial work on the base of maize seedlings fertilization, water injection, spraying at this stage has not been applicable, this will waste a lot of manpower and time, low efficiency, great workload. The image processing method is used to locate the base of maize seedling, to realize intelligent field operation, to apply fertilizer to the base of maize seedling intelligently, and to inject water into the base of maize seedling. Spraying medicine has become a trend in the future. On the basis of reading a large number of domestic and foreign literature, this paper studies the design method and operation flow of the image location method of maize seedling base. It provides feasibility for intelligent field operation in the future. This paper focuses on the following aspects: 1). The factors that affect the quality of image acquisition and image denoising and grayscale processing are studied experimentally. The height of image acquisition is analyzed through the image acquisition system. On the basis of the research of angle and illumination, the three elements of the relatively clear still image were studied. On the basis of collecting the clear image, the 3 脳 3 mean filter method was selected to remove the noise of the maize seedling image. In order to ensure the speed of follow-up processing as fast and accurate as possible. Finally, the weighted average method is used to grayscale the maize seedling image. (2) on the basis of gray image, the main trunk is extracted from the image segmentation and connected area. In order to distinguish the target information from the background information more obviously, the Ostu threshold method is chosen to segment the image. In order to make the target information more concise and convenient for subsequent processing. In this paper, the edge detection method of the segmented binary image is studied in order to make the target information clear and complete. In this paper, the Canny operator is used to detect the edge of the image, and the edge image is overlapped with the binary image again for morphological processing, resulting in the final clarity. Complete binary figure. (4) the methods of maize seedling base location, growth angle, stem diameter and stem identification were studied. The comparison of several localization methods was carried out. The rotation projection positioning method was selected to locate the base of maize seedling. The growth angle was extracted on the basis of image rotation, and the stem diameter and stem were recognized.
【学位授予单位】：沈阳农业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;S513

【参考文献】