This article aims to show an approach for impact crater extraction based on normal cycles to resolve issues associated with the insufficient and inaccurate extraction of quantitative information about Martian impact craters. We propose an algorithm combining topology information and convolutional neural network noise filtration of the MOLA data sample. The proposed method obtains better results than other algorithms.

• Hristo Hristov
• Nicole Christoff

2022 57th International Scientific Conference on Information, Communication and Energy Systems and Technologies (ICEST)

https://ieeexplore.ieee.org/document/9828781