基于Kinect的农作物长势深度图像实时获取算法  

Real-time obtaining algorithm for the range image of crop growth based on Kinect

作　　者：江晓庆[1,2] 肖德琴[2] 张波[1] 陈剑[2] 

机构地区：[1]华南农业大学现代教育技术中心,广东广州510642 [2]华南农业大学信息学院,广东广州510642

出　　处：《广东农业科学》2012年第23期195-199,F0003,共6页Guangdong Agricultural Sciences

基　　金：国家"十二五"农村领域国家科技计划项目(2011BAD21B01-1);NSFC-广东联合基金重点项目(U0931001)

摘　　要：快速、准确、实时获取农作物长势形态图像信息已经成为数字农业生产管理的趋势和必要手段,而深度图像信息包含彩色图像所没有的三维信息,可为农作物的生长模型三维重建以及实时长势形态监测提供有价值的数据。本研究采用微软公司开发的一款廉价体感游戏设备Kinect,尝试了对农作物长势形态深度图像进行实时监测研究,在介绍Kinect深度成像原理的基础上,提出了采用Kinect获取农作物长势深度图像的算法以及提取3D点云世界坐标的算法,并开展了初步试验。试验结果表明,该方法可以获取良好的三维图像数据信息,能够为后期农作物的生长模型三维重建、长势状态分析、病虫害实时监测等研究提供有价值的数据依据。Real-time obtaining the image information of crop growth quickly and accurately has become the tremt and neeessat.w means of digital agriculture, while the range image eontaining the three-dimensinnal information which is not included in the olor image can provide valuable data for the three-dimensional reconstruction of crop growth model and real-time growing mnnitoring. In this paper, we tried using a cheap game device alled Kineet developed by the Microsoft for real-time monitoring the range imge of rop growth. On the basis of analyzing Kineet depth imaging, this paper proposes the algorithm for obtaining the range image of crop growth ilh Kinet and the algorithm for getting 3D point cloud world tnordinates, then carried out the preliminary experiments. Experimental results showed that the method could obtain good three-dimensional image information so as to provide valuable data for three-dimensional reconstruction of erop growth model, growing state analysis, real-time pest monitoring and so on.

关 键 词：农作物长势 深度图像 3D点云坐标 KINECT 

分 类 号：S121[农业科学—农业基础科学]