Processing images taken by drones for determining the treedensity in park and cities gardens

One of the most important aspects in which modern technologies should be introduced is the environmental dimension. The environment we live in includes the protection of its resources, and the most important of these resources is trees. Major advances in image recognition in computer vision and machine learning algorithms aim to provide tools to assist forest management. Intelligent tools can be provided for processing images captured by the drone to identify the crowns of trees. High biodiversity requires strategies to conserve and manage natural resources, both individually and across the region. In this article, we use machine learning and focus on developing an automated approach for tree detection and classification. The system has been applied to photographs of date palm forests and their surroundings in Saladin, Iraq. High resolution images were used as dataset to train and test the system. While the crown detection accuracy for palm trees reaches 92%, tests are carried out to determine the cultivated areas. Detection and characterization of individual tree crowns is the first step in developing a forest surveillance approach for palm trees. The state of forests can be described and a descriptive analysis provided as an output of our proposed system