Texture Based Weed Detection Using Multi Resolution Combined Statistical and Spatial Frequency (MRCSF)

Authors: R.S.Sabeenian, V.Palanisamy

Abstract:

Texture classification is a trendy and a catchy technology in the field of texture analysis. Textures, the repeated patterns, have different frequency components along different orientations. Our work is based on Texture Classification and its applications. It finds its applications in various fields like Medical Image Classification, Computer Vision, Remote Sensing, Agricultural Field, and Textile Industry. Weed control has a major effect on agriculture. A large amount of herbicide has been used for controlling weeds in agriculture fields, lawns, golf courses, sport fields, etc. Random spraying of herbicides does not meet the exact requirement of the field. Certain areas in field have more weed patches than estimated. So, we need a visual system that can discriminate weeds from the field image which will reduce or even eliminate the amount of herbicide used. This would allow farmers to not use any herbicides or only apply them where they are needed. A machine vision precision automated weed control system could reduce the usage of chemicals in crop fields. In this paper, an intelligent system for automatic weeding strategy Multi Resolution Combined Statistical & spatial Frequency is used to discriminate the weeds from the crops and to classify them as narrow, little and broad weeds.

Keywords: crop weed discrimination, MRCSF, MRFM, Weeddetection, Spatial Frequency.

Digital Object Identifier (DOI): doi.org/10.5281/zenodo.1332610

References:

[1] Irshad Ahmad, Abdul Muhamin Naeem, Muhammad Islam, and Shahid Nawaz, "Weed Classification using Histogram Maxima with Threshold for Selective Herbicide", Applications proceedings of world academy of science, engineering and technology volume 21 January 2007.
[2] Irshad Ahmad, Abdul Muhamin Naeem, and Muhammad Islam, "Real- Time Specific Weed Recognition System Using Histogram Analysis", proceedings of world academy of science, engineering and technology volume 16 November 2006.
[3] P. M. Graniatto, H. D. Navone, P .F. Verdes, and H.A. Ceccatto, "Weed Seeds Identification By Machine Vision", Computers and Electronics in agriculture 33, 2002, 91-103.
[4] Gerrit Polder, Frits K. van Evert, Arjan Lamaker, Arjan de Jong, Gerie van der Heijden, Lambertus A.P.Lotz, Ton van der Zalm, Corné Kempenaar "Weed Detection Using Textural Image Analysis".
[5] R.S.Sabeenian, V.Palanisamy, ÔÇÿComparision of Efficiency for Texture Image Classification using MRMRF & GLCM Techniques- Published in the International Journal of Computers Information Technology and Engineering (IJCITAE) 2, 2008, 87-93.
[6] Haralick, R.M, ÔÇÿStatistical and structural approaches to Texture-IEEE proceedings.67, 1979,pp 786-804.
[7] Lei Wang, Jun Liu, ÔÇÿTexture classification using Multiresolution Markov Random Field Model-, Pattern Recognition Letters 20, 1999, pp 171-182.
[8] Philip.Brodarz, ÔÇÿTexture A Photographic album for artists and designers- New York, Rein Hold,1968.
[9] J.Pérez,F.L├│pez,J.V.Benlloch, S.Christensen,"Colour and Shape Analysis Techniques For Weed Detection in Cereal Fields".First European Conference for Information Technology in Agriculture, Copenhagen, 15-18 June, 1997.
[10] Arivazhagan.S, Ganesan.L, ÔÇÿTexture Classification using Wavelet Transform-, Pattern Recognition Letters.24, 2003,1513-1521
[11] Wang, L., Liu, J., Li, S.Z.(1998), ÔÇÿTexture Classification using wavelet decomposition with Markov random field models-, In the Proceedings of International Conference on Pattern Recognition, 1998.