African Journal of
Agricultural Research

  • Abbreviation: Afr. J. Agric. Res.
  • Language: English
  • ISSN: 1991-637X
  • DOI: 10.5897/AJAR
  • Start Year: 2006
  • Published Articles: 6576

Full Length Research Paper

Bayesian discriminant analysis of plant leaf hyperspectral reflectance for identification of weeds from cabbages

Wei Deng
  • Wei Deng
  • Beijing Research Center of Intelligent Equipment for Agriculture, Beijing 100097, China.
  • Google Scholar
Yanbo Huang
  • Yanbo Huang
  • United States Department of Agriculture, Agricultural Research Service, Crop Production Systems Research Unit, Stoneville, Mississippi, USA.
  • Google Scholar
Chunjiang Zhao
  • Chunjiang Zhao
  • Beijing Research Center of Intelligent Equipment for Agriculture, Beijing 100097, China.
  • Google Scholar
Liping Chen
  • Liping Chen
  • Beijing Research Center of Intelligent Equipment for Agriculture, Beijing 100097, China.
  • Google Scholar
Xiu Wang
  • Xiu Wang
  • Beijing Research Center of Intelligent Equipment for Agriculture, Beijing 100097, China.
  • Google Scholar


  •  Received: 16 September 2015
  •  Accepted: 19 January 2016
  •  Published: 18 February 2016

Abstract

In order to spray herbicides accurately on targets, this study focused on spectral classification of weeds and crops for potential to rapidly detect weeds in crop fields. A 350 ~ 2500 nm FieldSpec-FR spectroradiometer was used to measure spectral responses of the canopies of the seedling vegetables, cabbage ‘8398’ and cabbage ‘Zhonggan 11’, and weeds, Barnyard grass, green foxtail, goosegrass, crabgrass, and Chenopodium quinoa, at five- and seven-week growth stages (WGS). First, the characteristic wavelengths (CW) were determined using Principal Component Analysis (PCA). Then, the plants were classified using Bayesian discriminant analysis with the reflectance of the CWs. The results of spectral analysis indicated that the different growth stages of cabbages had little influence on the spectral identification of cabbages and weeds. The eight CWs determined were used as the input to the model for Bayesian discriminant analysis to classify two varieties of cabbages and five weeds with the correct classification rate of 84.3% for model testing. When the two varieties of cabbages were considered as the same category, the correct classification rate was improved to 100%. It was concluded that Bayesian discriminant analysis could be used to identify weeds from seedling cabbages using leaf hyperspectral reflectance.

 

Key words: Weed identification, spectrum analysis, visible and near-infrared, Bayesian discriminant, seedling weed, seedling cabbage.