基于Kinect的温室番茄盆栽茎干检测与分割  

Detection and Segmentation of Stems of Greenhouse Potted Tomato Based on Kinect

作　　者：付代昌 徐立鸿[1] 李大威[1] 辛龙娇 

机构地区：[1]同济大学电子与信息工程学院,上海201804

出　　处：《现代农业科技》2014年第3期336-338,343,共4页Modern Agricultural Science and Technology

基　　金："十二五"国家863计划重点课题(2013AA100305);国家自然科学基金项目(61174090)

摘　　要：立体视觉系统能够得到图像的深度信息,可以很好地将植物前景与温室的背景分开。茎干是植物的骨架,也是识别植株的重要特征。通过采用基于Kinect的立体视觉系统对温室盆栽植物的茎干进行检测和分割。通过细化算法、Hough变换、基于直方图的统计方法及Otsu算法在深度图像上进行植物茎干检测,从而分割植物的茎干。最后基于OpenGL图形平台,实现了番茄植株个体的形态可视化,从而为番茄植株生长的可视化奠定了技术基础。Stereo vision system can obtmn the depth information of an image, so it could well separate the plant region from complicated green- house backgounds.Stems are the skeletons of plants, and they are important characteristics of the recognition of plants.Kinect stereo vision system was taken advantage of carrying out the detection and segmentation of stems of greenhouse potted plant.In this pape,it＇s detected stems in depth images via thinning algorithm, Hough transformation, statistical methods based on histogram and 0tsu algorithm, and eventually segmented the stems.In the end, the visualization of the tomato plant was realized based on the OpenGL graphic platform,consequently laying the technological foundation of the realization of tomatoes＇ growth.

关 键 词：KINECT 细化算法 OTSU算法 温室番茄盆栽 茎干分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]