A Green Vegetation Extraction Based-RGB Space in Natural Sunlight

Abstract

Green vegetation segmentation in color images is a fundamental issue for automated remote sensing and machine vision applications, plant ecological assessments, precision crop management, and weed control. A simple green vegetation feature extraction method (GVFE) is proposed in this paper to segment the green vegetation from their non-green backgrounds due to the fact that the green component content is always greater than that of the red and blue in RGB color space. The conventional based-auto-threshold method, ExG (Excess Green) was compared with GVFE, in which a green index ratio was defined to evaluate the performance of them. A digital color image set of single Canna flower taken in natural lighting were used to test them. Experimental results have showed that GVFE has superior performance over ExG+auto-threshold in term of stability, and is insensible to illuminant variations.

A Semi-Automated Method to Extract Green and Non-Photosynthetic Vegetation Cover from RGB Images in Mixed Grasslands

“…Because field-taken digital RGB images with high spatial resolution potentially provide more accurate estimation of vegetation cover than visual methods by reducing the impact of human subjectivity [1,14]. They have been widely used for estimating forest canopy cover i.e., gap fraction analysis, [17,18], forest understory cover [16], crop cover [4,[19][20][21][22], crop residual cover [23] and grassland vegetation cover [5,24].…”