作 者：
刘双喜;李友永;荆林龙;王富;孙林林;王金星
关键词：
苹果树叶片;含氮量检测;色彩恒常理论;色调差
摘 要：
该文旨在应用数字图像处理技术建立苹果树叶片氮含量检测模型。该方法以自然光条件下的叶片图像为分析对象,首先通过色彩恒常理论MSR算法进行去光照处理,然后应用阈值分割、改进的Sobel算子剔除叶片噪声点、叶脉等的影响,最后对不同含氮量的叶片与潘通标准色卡进行色调差分析,得到果树叶片色调差和含氮量之间的关系,建立出苹果树叶片的含氮量检测模型。经试验验证,苹果树叶片去光照图像与标准色卡图像的色调差与果树各个时期的叶片含氮量的相关系数均在0.75以上,且回归方程的F检验P值均小于0.01,具有极显著的相关关系,其中盛花期传统施肥组、结果中期传统施肥组、施复合肥组、采收期传统施肥组相关系数均在0.8以上,用最小二乘法建立的检测模型的检测值与实际值的相对误差小于10%。说明利用色调差测定苹果叶片含氮量的方法可行,该方法为实时、准确地监控苹果树整个生长过程中叶片含氮量的变化以及合理地施用氮肥提供科学理论基础。
译 名：
Rapid detection of nitrogen content in leaves of apple trees based on hue-difference natural lighting
作 者：
Liu Shuangxi;Li Youyong;Jing Linlong;Wang Fu;Sun Linlin;Wang Jinxing;Shandong Provincial Key Laboratory of Horticultural Machineries and Equipments;College of Mechanical and Electronic Engineering,Shandong Agricultural University;
单 位：
Liu Shuangxi%Li Youyong%Jing Linlong%Wang Fu%Sun Linlin%Wang Jinxing%Shandong Provincial Key Laboratory of Horticultural Machineries and Equipments%College of Mechanical and Electronic Engineering,Shandong Agricultural University
关键词：
apple tree leaves;;nitrogen content detection;;color constant theory;;hue difference
摘 要：
This paper was aimed to apply digital image processing technology to establish a model of nitrogen content detection for apple trees.This method was took leaf images under natural light conditions as the analysis object.Firstly,it was performed the delighting process using the color constancy theory MSR algorithm,and then used threshold segmentation,improved Sobel operator to eliminate the influence of blade noise points,veins,etc.,and finally to different nitrogen content.The amount of leaves and the Pantone standard color card were used for the analysis of the tonal difference,and the relationship between the color tone difference and the nitrogen content of the fruit trees was obtained,and a nitrogen content detection model for apple tree leaves was established.Through experiments,the correlation coefficient between the hue difference between the apple deforested leaves image and the standard chromatographic image and the leaf nitrogen content of fruit trees in each period was above 0.75,and the F-test Pvalue of the regression equation was less than 0.01,which was extremely significant.The correlations among the traditional flower fertilization group,the middle-term traditional fertilization group,the compound fertilization group,and the traditional fertilization group during the harvest period all had correlation coefficients above 0.8.The detection values and actual values of the detection model were established by the least squares method.The relative error was less than 10%.The method of determining the nitrogen content of apple leaves by using the hue difference is feasible.This method provides a scientific theoretical basis for real-time and accurate monitoring of the changes of nitrogen content in the leaves of apple trees and reasonable application of nitrogen fertilizers.
