A recent study presented at the 56th Lunar and Planetary Science Conference (LPSC) explores the use of artificial intelligence (AI) to enhance mapping and imaging on Mars. The research focuses on improving orbital images from the Mars Reconnaissance Orbiter (MRO) Context Camera (CTX) using machine learning models.
Dr. Andrew Annex from the SETI Institute led the study, aiming to accelerate scientific discovery and maximize the value of existing Mars datasets. He developed a visual search engine capable of analyzing global CTX mosaic images, identifying specific image similarities across the planet.
The study evaluated content-based image retrieval (CBIR), OpenAI CLIP, and cloud computing architecture. CBIR scans databases for similar images based on content, while OpenAI CLIP compares images and text using large datasets. Cloud computing manages extensive data through remote servers.
Dr. Annex successfully used machine learning to analyze global CTX mosaic images on Mars, including searching and identifying specific image similarities across the Red Planet. This research opens doors for improvements in search queries on planetary surfaces across the solar system.
Since NASA's Mariner 4 captured the first image from a Mars orbiter in 1965, numerous Mars orbiters have provided detailed images of the planet's surface. The entire surface of Mars has been imaged by NASA's Context Camera and High Resolution Imaging Science Experiment (HiRISE) camera.
Dr. Annex emphasizes the importance of machine learning to improve image analysis methods, noting that while computing power has increased, the speed of data analysis has not kept pace. Machine learning offers flexibility and speed in automating tasks, complementing existing methods.
Machine learning is a tool that can complement and enhance existing methods and analysis, improving the speed and accuracy of image analysis, leading to new discoveries about Mars and other planets.