Deep Learning Based Microatoll Detection From Drone Images

Visual object detection has made significant progress with the advent of deep neural networks and has been extensively applied. This thesis reports a novel application that aims to detect individual microatolls, which are circular coral colonies, from island images captured by drones. We first describe data collection and labelling to create a novel microatoll dataset for the microatoll detection task from drone images. Upon this dataset, the state-of-the-art object detectors are then evaluated for this task. To better integrate a detector with the characteristic of microatolls, we propose a modified detector called Microatoll-Net. It actively extracts features from the surrounding area of a microatoll to differentiate it from distractors to improve detection. Multiple ways to incorporate this information into the detector are designed. The experimental study shows the efficacy of the proposed Microatoll-Net, especially on the most challenging area for detection. Besides, in geographical research, the position of a microatoll is more important than its size. It means that we shall pay more attention to detecting the centre of a microatoll instead of generating its bounding box. Motivated by this, we transform this object detection task into an object centre detection task.