Tree Census Using Circular Hough Transform and GRVI

Abstract

Abstract Remote sensing techniques provide a potentially low-cost alternative to field-based assessments like tree census, tree crown and tree type identification. But it requires the development of algorithms for accurate extraction of the required information. The data generated out of Multispectral imagery (MSI) with high-spatial-resolution systems provides an opportunity for analysis and mapping of the tree crowns. Tree crown mapping with MSI is generally done by the use of the standard Normalized Difference Vegetation Index (NDVI). MSI can also be taken from Google Earth. Google Earth images have only the red, green and blue band imagery. The response of vegetation can also be reliably detected by means of a simple threshold, using the Green Red Vegetation Index (GRVI). Our method consists of the combination of range detection for GRVI and Circular Hough Transform (CHT) for tree crown extraction. The crown identification also helps in the tree count. Our experiments reveal that the proposed approach yields state-of-the-art tree counts for different datasets. For the first dataset the completeness and the correctness are found to be 0.81 and 0.95. The completeness and correctness are found to be 0.97 and 0.84 respectively for the second dataset.