相机标定对玉米自动化观测评估结果的影响  

Effects of Camera Calibration on Maize Automatic Observation Results

作　　者：张志红[1] 余国河[2] 田东哲 胡锦涛[2] 杨玉辉 Zhang Zhihong;Yu Guohe;Tian Dongzhe;Hu Jintao;Yang Yuhui(CMA·Henan Key Laboratory of Agrometeorological Support and Applied Technique,Zhengzhou 450003,China;The 27thResearch Institute of China Electronic Technology Group Corporation;Inner Mongolia Tuquan County Meteorological Bureau)

机构地区：[1]中国气象局河南省农业气象保障与应用技术重点开放实验室,郑州450003 [2]中国电子科技集团公司第二十七研究所 [3]内蒙古突泉县气象局

出　　处：《单片机与嵌入式系统应用》2019年第4期26-29,共4页Microcontrollers & Embedded Systems

基　　金：中国气象局农业气象保障与应用技术重点开放实验室开放研究基金(AMF201607)

摘　　要：作物气象自动观测需要采用数码相机采集作物实景图像,进而利用图像处理算法对作物特征进行提取。受相机畸变参数影响,实景图像存在不同程度的失真,导致观测区域面积、密度、发育期等识别结果有所偏差。本文以玉米为例,根据玉米发育期自动观测评估结果对相机的畸变参数进行分析,通过对相机标定算法的研究,利用Matlab工具箱对相机进行标定,提高对玉米发育期自动识别的准确性。The automatic crop meteorological observation requires the actual crop image collected by the digital camera,and the crop features are extracted by the image processing algorithm.The real image has different degrees of distortion because of the camera distortion parameters,which results in some errors in the recognition results of area,density and developmental period.In this work,maize is taking as an example.According to the results of the automatic observation and evaluation during the corn development,the camera distortion parameters are analyzed.Through the study of camera calibration algorithm,the camera is calibrated with Matlab toolbox.As a result, the accuracy of automatic identification during the corn development is improved.

关 键 词：图像处理 相机标定 MATLAB 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]