Up to journal
for Research & Developement
Journal of Materials Science: Materials in Electronics
Impact Factor
0/000
Available
2000 - 2020
Volumes
8
Issues
25
Articles
218
Open Access
66


Journal of Researches in Mechanics of Agricultural Machinery
Providing a machine vision system based on meta-heuristic classifier to classify two weed types


 submission: - | acception: - | publication: 08/10/2019

DOI 

Authors
Sajad Sabzi1, Hosein Javadikia2, Yousef Abbaspour-Gilandeh3*

1-،sajadsabzi2@gmail.com

2-،h.javadikia@razi.ac.ir

3-،abbaspour@uma.ac.ir



469 Downloads

Abstract

Site-specific spraying is a new method to weed exclusion. For this reason, a machine vision system is suggested in this study based on video processing and meta-heuristic classifier for online identification of potato plant and two types of weeds PortulacaOleracea, and Xanthium strumarium. In order to train the suggested machine vision system, ۵ hectares of lands of Marfona potato in Kermanshah province were selected for filming. After extracting some features in fields of color features and spectral descriptors of texture using two methods of hybrid artificial neural network - ant colony (ANN-ACO) and hybrid artificial neural network - Particle swarm optimization (ANN-PSO), the effective features were selected. Finally, by analyzing the results of selected features set by each of the two mentioned methods, the selected features were selected using method of artificial ANN-PSO in order to use in machine vision system. Selected features were color index for vegetation cover in YCbCr color space, the extra second component index in HSV color space, the first and second component contrast in HSV color space, the mean of first component in HSI color space and the mean of S(r). The results of classification showed that method of hybrid artificial neural network - Differential Evolution (ANN-DE) with high accuracy of ۹۹% is able to identify potato plant and two different types of weeds.




Keywords

Weeds Site specific spraying Classification Machine vision 



Download fulltext PDF


Open Access

Download

Citation





References





Announcements  

  •  No announces available
Indexes  

 

no item exists