Automated Detection of Craters and Boulders Using Machine Learning on OHRC Images

Detection of Craters and Boulders is a critical task in Space. In earlier times, Counting was performed manually on the images. This Counting takes a lot of time, and the probability of Human error is high. This research aims to automate this process and make it easier for Humans. Making this process fully automated requires state-of-the-art machine-learning algorithms. These algorithms have much efficiency and precision, which results in fewer errors. The help of an Orbiter High-Resolution Camera (OHRC) and Computer Vision techniques make data more accurate for processing. Later, this manuscript applies certain machine learning algorithms such as Support Vector Machines (SVM), Random Forest, and Convolutional Neural Networks (CNN) on these data to detect features such as rocks, craters, and boulders with promising accuracy. The author observes that the Support Vector Machine results (SVM) have a better level of precision. Additionally, this research identifies the most effective algorithm for crater and boulder detection.