CNN Based Technique for Automatic Tree Counting Using Very High Resolution Data

摘要

Coconut is one of the economically grown crops in India. In this paper, we develop an automatic method for counting the number of coconut trees in UAV images. The availability of high resolution remote sensing images helps people in having large amounts of detailed digital imaging of vegetation areas. Today, the estimated coconut tree count can be determined in a short duration of time through high resolution drone images with low cost and labor. The goal is to find new methods to determine coconut trees using remote sensing. Deep learning techniques with convolutional neural network (CNN) algorithms is used to detect the coconut trees.