Vol. 7, Issue 6 (2019)
An overview on crop-weed discrimination based on digital image processing using textural features
Author(s): Pankaj Malkani, Atish Sagar, Asha KR, Abhinav Dubey and Prashant Singh
Abstract: Weed control is a serious issue for maximizing the yield. The traditional weed management approaches are less effective, and requires high labor force in peak seasons. Modern machineries using site-specific weed management (SSWM) system manages weed precisely and delivered a precise amount of chemicals to only weeds. The heart of the SSWM system is a digital image processing system that involves image preprocessing, vegetation segmentation, feature extraction, and classification. The feature extracted from the images using color, spectral and spatial method had several limitations concerning color; inter-row spacing and crop-weed growth at different stages. The texture-based feature extraction process for image classification is the best possible way over others. It involves statistical GLCM matrix, structure, wavelet, and Model-based textural features. Statistical and wavelet based textural features are most commonly used for crop-weed classification system. The texture features can recognize weed with more than 90% accuracy and are more effective than other feature extraction methods. Therefore, it had huge potential and scope in SSWM machinery.
Pages: 2514-2520 | 221 Views 61 Downloads
download (7006KB)
How to cite this article:
Pankaj Malkani, Atish Sagar, Asha KR, Abhinav Dubey, Prashant Singh. An overview on crop-weed discrimination based on digital image processing using textural features. Int J Chem Stud 2019;7(6):2514-2520.
