基于光谱特性的植保机智能图像识别研究  

Research on Intelligent Image Recognition of Plant Protection Machine Based on Spectral Characteristics

作　　者：王秀珍[1] Wang Xiuzhen(Zhengzhou University of Science and Technology,Zhengzhou 450064,China)

机构地区：[1]郑州科技学院,郑州450064

出　　处：《农机化研究》2023年第10期207-210,共4页Journal of Agricultural Mechanization Research

基　　金：河南省科技厅软科学项目(182400410595)。

摘　　要：杂草图像识别是植保机作业过程中的关键环节,而利用光谱特性分析技术可以快速对种植区域的杂草进行识别。为此,利用ASD光谱分析仪连续进行5组光谱数据采集,经分析处理后得出光谱数据;按照最小值原则选取595、710、755、950nm共4个特征波长,建立典型判别函数模型和贝叶斯判别函数模型,对预测集合数据进行预测。结果表明:贝叶斯判别函数模型能够较好地进行种植区域杂草图像识别,且模型具有较高的稳定性,为植保机杂草图像识别探测传感器的开发提供了参考依据。Weed image recognition is the key link in the operation process of plant protection machine.The weeds in the planting area can be recognized quickly by using spectral characteristic analysis technology.In this paper,five groups of spectral data are collected continuously by ASD spectral analyzer.After analysis and processing,the spectral data are obtained.According to the minimum principle,four characteristic wavelengths of 595nm,710nm,755nm and 950nm are selected to establish typical discriminant function model and Bayesian discriminant function model to predict the prediction set data.The results show that the Bayesian discriminant function model can better recognize the weed image in the planting area,and the model has high stability,which provides a reference basis for the development of weed image recognition and detection sensor of plant protection machine.

关 键 词：植保机 图像识别 光谱特性 判别函数 

分 类 号：S224.3[农业科学—农业机械化工程] TP391.41[农业科学—农业工程]