节点文献
基于计算机视觉的水稻叶片信息研究
Research on the Leavers Information of Rice Based on Computer Vision
【作者】 童钊；
【导师】 廖桂平；
【作者基本信息】 湖南农业大学 ， 农业信息化技术， 2010， 硕士
【摘要】 植物叶片是植物进行光合作用和蒸腾作用的主要器官,叶片信息直接反映了作物的生长状况,这些信息对于农作物的生长监控具有非常重要的作用。本文以精准农业的发展趋势为指导,应用计算机视觉技术,希望建立更为准确、快捷的叶面积测量方法和利用叶片颜色特征值的特点结合实测的SPAD(Soil Plant Analysis Development)值建立基于计算机视觉的叶片颜色和SPAD值的预测模型。基于以上目的,本论文研究的主要内容包括：对水稻单叶片的面积进行测量,在使用传统的计算机视觉技术测量面积时发现使用叶片背面作为拍照面的测量效果好于使用正面的效果。且在此基础上,探索了两种基于计算机视觉法测量叶片面积的新方法；分析并解决了在大田环境中影响计算机视觉获取图像和测量方法的若干个关键问题,使得在使用计算机视觉对水稻叶片进行SPAD值预测时,精确度更高；在以上结论的基础上分别建立了实验室环境下和大田环境下水稻叶片颜色特征值与SPAD值的模拟模型,并预测其SPAD值,其精度最高可分别达到：97.5%和96.59%。本文的主要创新点在于：提出了基于计算机视觉的水稻叶片叶色网格分析法,能将水稻叶片表面分割成若干个区域,分析各个区域的颜色特征为使用SPAD仪测量位点选择提供了理论依据。
【Abstract】 Leaf is the main organ for plant photosynthesis and transpiration. Blade information directly reflects the growth status of crops, which is very important to monitor the growth of crops. In this paper, the trend of precision agriculture guidance using computer vision technology, hoping to establish a more accurate and efficient measurement of leaf area and use of leaf color characteristic value of the features of SPAD(Soil Plant Analysis Development) values established by actual computer vision-based color and SPAD values of leaves prediction model. Based on the above purpose, the main contents of this paper include:In the process of measuring the rice leaf area, it’s better to measure the leaf back than the leaf surface by using traditional computer vision techniques. Based on the above research, new method in measuring rice leaf area of computer vision measurement is explored; To analyze and solve the impact of the environment in field environment computer vision to obtain images and measurements of a number of key issues, making the use of computer vision, SPAD values of rice leaves to predict when the greater accuracy; In the above conclusions were established on the basis of a laboratory environment and field environment, the color characteristics of rice leaf SPAD value and the value of a simulation model and predict the SPAD value, and its accuracy is up to, respectively:97.5% and 96.59%;The main innovation of this paper is propose the rice leaf color grid analysis based on computer vision, which can divide the rice leaf surface into several regions, and analyze the color characteristics of each region, supplying the theoretical basis to the selection of SPAD sites.