基于最小阈值法的棉花幼苗识别研究

Study on Cotton Seedling Separation Based on Minimum Threshold Method

面对纯机械式分苗机构经常出现的漏分和多分的问题,本文提出使用基于最小阈值法的图像识别技术对棉花幼苗进行识别和分离。使用了HSV颜色空间内与光照条件无关的H值将RGB真彩色图像转化为灰度图像,以提高图像处理的实时性。基于最小值阈值法,提出一种能够有效地将处于复杂背景下的棉花幼苗茎部图像提取的方法,使用棉花幼苗茎部图像的面积和长宽比等几何信息能有效识别棉花幼苗,并得到棉花幼苗茎部的重心坐标。本文的研究证明了用图像处理中的最小阈值法识别棉花幼苗是可行的。

Image recognition technology based on the minimum threshold method has been applied in the cotton seedling recog- nition and separation to resolve the problem that could not exactly pick up one cotton seedling appearing in the manual cotton seedling sorting equipment. The H value in the HSV color space, having nothing to do with light conditions, was used to convert the RGB true color image into the corresponding gray image to improve the real-time ability of image processing. A method of threshold extraction is put forward for dividing the cotton seedling stem from complex background based on the minimum threshold value method. The cotton seedling can be recognized using the area and long-wide ratio of the stem＇s geometry infor- mation, and the core coordinates of the stem can be got correspondingly. Thus it is proved that the method of recognizing cotton seedling using minimum threshold method in the image processing is feasible.