Estimating the Speed of the ISS Using Computer Vision
The article describes a project that estimates the speed of the International Space Station (ISS) using only images of Earth, leveraging Python and OpenCV for computer vision techniques.
Why it matters
This project demonstrates the potential of using computer vision techniques to estimate the speed of objects in space, which could have applications in satellite tracking and space exploration.
Key Points
- 1Captured two images of Earth from the ISS
- 2Detected keypoints using SIFT and matched them using FLANN
- 3Measured pixel displacement and converted to real-world distance
- 4Calculated the speed based on the time difference between the two images
- 5Achieved an estimated speed of ~7.47 km/s, compared to the actual ISS speed of ~7.66 km/s
Details
The project uses a computer vision pipeline to estimate the speed of the ISS. It captures two images of Earth, detects keypoints using the SIFT algorithm, and matches them using the FLANN (Fast Library for Approximate Nearest Neighbors) method. The pixel displacement is then measured and converted to real-world distance using the Ground Sampling Distance (GSD). Finally, the speed is calculated based on the time difference between the two images. The author achieved an estimated speed of ~7.47 km/s, which is within a 2-3% difference from the actual ISS speed of ~7.66 km/s. The author notes that the original runtime images are lost, limiting the ability to reproduce the results, and suggests improvements such as better filtering of bad matches, a cleaner pipeline, and more robust handling of outliers.
No comments yet
Be the first to comment