Rock Hunting With Martian Machine Vision
arXiv:2104.04359
Abstract
The Mars Perseverance rover applies computer vision for navigation and hazard avoidance. The challenge to do onboard object recognition highlights the need for low-power, customized training, often including low-contrast backgrounds. We investigate deep learning methods for the classification and detection of Martian rocks. We report greater than 97% accuracy for binary classifications (rock vs. rover). We fine-tune a detector to render geo-located bounding boxes while counting rocks. For these models to run on microcontrollers, we shrink and quantize the neural networks' weights and demonstrate a low-power rock hunter with faster frame rates (1 frame per second) but lower accuracy (37%).
References in corpus (6)
- YOLOv4: Optimal Speed and Accuracy of Object Detection
- Scaled-YOLOv4: Scaling Cross Stage Partial Network
- TinyTL: Reduce Activations, Not Trainable Parameters for Efficient On-Device Learning
- Machine Learning Based Path Planning for Improved Rover Navigation (Pre-Print Version)
- Overhead MNIST: A Benchmark Satellite Dataset
- The PIXL Instrument on the Mars 2020 Perseverance Rover