UAV multispectral image for tree crown delineation of mangifera indica using object based image analysis [OBIA] / Nik Syaza Sahiera Nik Nazmy

Tree crown delineation are the basis of precision forestry and agriculture precision forestry and agriculture rely on accurate tree delineation. There are many techniques and algorithms that can be applied in order to create the tree crown delineation whether in forest or plantation. However, the result of each segmentation technique or algorithm used derive difference result. This study was conducted at Harumanis plantation located at UiTM Arau, Perlis with the aim of to access the tree crown delineation of Harumanis plantation from UAV multispectral using two algorithms between watershed segmentation and local maxima and local minima. In this study, conventional method was used to derive the parameter and positioned of tree, functioned as reference. The UAV multispectral image was used to determine the individual tree crown delineation by using algorithm of watershed and local maxima and minima which had been processed using eCognition software. Besides that, the manually digitized of tree crown delineation was conducted to perform the accuracy assessment. The output of tree crown delineation of watershed and local maxima and minima shown watershed algorithm has higher accuracy with 83.3% of Total 1:1 match and 73% of goodness of fit while local maxima and minima are 81.7% and 71%. This research will benefit the forest management for the future planning and forest quantification estimation using remote sensing technology.