摘要
在农业生产中,杂草会影响农作物的生长。为了实现杂草处理自动化,需要采用基于机器视觉的杂草识别技术帮助机器人快速、准确地识别杂草,从而精确使用除草剂以提高药物的利用率、减少环境污染。通过文献研究,文章分析了提高杂草识别率的图像预处理技术和基于人工神经网络的模式识别算法,提出了杂草识别技术存在的问题及改进方法。
In agriculture, weeds can affect the growth of crops. In order to realize the process automation, the weed identification based on machine vision technology can help the robot quickly and accurately identify weeds, thus the robot can accurately use herbicides in order to improve the utilization of drugs and reduce environmental pollution. Based on the research, this paper analyzed the image preprocessing techniques which can enhance the rate of identification and pattern recognition algorithm based on artificial neural network, and present the weed identification method problems and improving methods.
引文
[1]SOGAARD H T,LUND I,GRALIA E.Real-time application of herbicides in seed lines by computer vision and microspray system[J].American Society of Agricultural and Biological Engineers,2006(12):118-127.
[2]尹建军,沈宝国,陈树人.基于机器视觉的田间杂草定位技术[J].农业机械学报,2010(6):163-166.
[3]王军敏,姬鹏飞.基于微分算子的边缘检测及其应用[J].平顶山学院学报,2015(2):85-89.
[4]李寒,基于机器视觉的目标检测在精细农业中的关键技术研究[D].北京:中国农业大学,2014.
[5]杨会清,孙丽燕,刘红.田间杂草识别的机器视觉系统及机械除草研究[J].农业机械,2011(28):126-129.
[6]YAN H.Skew correetion of document images using interline cross-correlation[J].Cvgip Graphical Models&Image Processing,1993(6):538-543.
[7]李东明.基于机器视觉的田间杂草识别方法研究[D].天津:河北工业大学,2008.
[8]黄锋.基于径向基函数神经网络的图像识别研究[D].太原:太原理工大学,2007.
[9]朱登胜,潘家志,何勇.基于光谱和神经网络模型的作物与杂草识别方法研究[J].光谱学与光谱分析,2008(5):1102-1106.
[10]卜富清,王茂芝,于庆刚.基于BP神经网络的数字识别[J].长江大学学报(自科科学版),2009(2):293-294.
[11]BURKS T F,SHEARER S A,HEATH J R,et al.Evaluation of neuralnetwork classifiers for weed species discrimination[J].Biosystems Engineering,2005(3):293-304.