红外传感器与机器视觉融合的果树害虫识别及计数方法

为了解决果园环境中单一的害虫监测技术存在的缺陷,该研究将红外传感器和机器视觉识别技术进行融合,从两个角度对目标害虫进行识别计数。选取梨小食心虫、苹小卷叶蛾、桃蛀螟、干扰物进行试验,通过实验室人工随机散落试验样本,获得其红外传感器以及机器视觉图像的识别结果,构造融合计数计算公式,通过计算得到害虫计数结果。结果显示:梨小食心虫、苹小卷叶蛾、桃蛀螟、干扰物的红外分类阈值分别为2.25、9.06、17.88、28.38,其红外识别范围分别为(0,5]、(5,13]、(13,23]、(23,32];梨小食心虫、苹小卷叶蛾、桃蛀螟、干扰物的红外识别准确率分别为92%、78%、80%、88%,图像识别准确率...

Full description

Saved in:
Bibliographic Details
Published in农业工程学报 Vol. 32; no. 20; pp. 195 - 201
Main Author 田冉 陈梅香 董大明 李文勇 矫雷子 王以忠 李明 孙传恒 杨信廷
Format Journal Article
LanguageChinese
Published 天津科技大学电子信息与自动化学院,天津 300222%国家农业信息化工程技术研究中心/农业部农业信息技术重点实验室/北京市农业物联网工程技术研究中心,北京,100097%天津科技大学电子信息与自动化学院,天津,300222 2016
国家农业信息化工程技术研究中心/农业部农业信息技术重点实验室/北京市农业物联网工程技术研究中心,北京 100097
Subjects
Online AccessGet full text
ISSN1002-6819
DOI10.11975/j.issn.1002-6819.2016.20.025

Cover

More Information
Summary:为了解决果园环境中单一的害虫监测技术存在的缺陷,该研究将红外传感器和机器视觉识别技术进行融合,从两个角度对目标害虫进行识别计数。选取梨小食心虫、苹小卷叶蛾、桃蛀螟、干扰物进行试验,通过实验室人工随机散落试验样本,获得其红外传感器以及机器视觉图像的识别结果,构造融合计数计算公式,通过计算得到害虫计数结果。结果显示:梨小食心虫、苹小卷叶蛾、桃蛀螟、干扰物的红外分类阈值分别为2.25、9.06、17.88、28.38,其红外识别范围分别为(0,5]、(5,13]、(13,23]、(23,32];梨小食心虫、苹小卷叶蛾、桃蛀螟、干扰物的红外识别准确率分别为92%、78%、80%、88%,图像识别准确率分别为92%、88%、92%、90%,融合计数精度分别为98%、92%、94%、96%。可见,将红外传感器和图像识别技术相融合能够提高果树性诱害虫的识别准确率,为果园害虫的合理防治提供参考。
Bibliography:machine vision; image recognition; data fusion; infrared sensor; fruit tree pests
Traditional single monitoring technique in orcha rd environment has such shortages as weak effectiveness, inaccurate count and pooruniversality. Now existing pest monitoring methods include acoustic measurement, piezoelectric measurement, infrared measurement and machine vision recognition technology. In view of this, the future development trend of pest detection technology will undoubtedly be a variety of detection methods combined with each other. Comprehensive utilization of the existing testing methods will form a multiple information fusion technique to detect and provide reliable scientific decision based on comprehensive prevention and control of fruit pests, and the loss will be reduced to a minimum. In this paper, infrared measurement and machine vision recognition technology are integrated to identify pest species and count pest populations, and information of pests is obtained from 2 aspects. The accuracy of the fusio
ISSN:1002-6819
DOI:10.11975/j.issn.1002-6819.2016.20.025