Image Segmentation Model for Resident Space Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting and tracking resident space objects (RSOs) are slow and require human operators, leading to potential collisions and inaccuracies in space situational awareness (SSA).
Innovation Solution
A method is developed to generate simulated images for training image segmentation models to detect RSOs, involving the calculation of imaging satellite coordinates, determination of the field of view, and simulation of RSO and star coordinates. The method includes adding noise and translating coordinates to pixel coordinates, and choosing an imaging mode and exposure time to render the simulated image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods are used for detecting and tracking RSOs, then human operators can analyze data, but the response time is slow and tracking accuracy is insufficient
Solution Approach 1:
The patent replaces the mechanical system of human operator analysis with an automated image segmentation model based on deep learning. The model processes optical images to detect and track RSOs autonomously, eliminating the time delay associated with manual analysis while improving measurement precision through algorithmic consistency and speed.
Solution Approach 2:
The system enables self-service by allowing the image segmentation model to autonomously detect, track, and classify RSOs without human intervention. The model processes images, identifies objects, and provides tracking information automatically, making the system self-sufficient in performing SSA tasks.
2Measurement precision
If simulated training images are generated using complex rendering processes, then model training accuracy improves, but computational resources and processing time increase
Solution Approach 1:
The patent applies preliminary action by pre-generating simulated training images with known RSO positions and characteristics before actual deployment. This allows the model to be trained in advance on diverse scenarios including edge cases, reducing the need for complex real-time processing and improving readiness while optimizing resource usage during operational phases.
Solution Approach 2:
The system creates copies of real-world scenarios through simulated training images that replicate various imaging conditions, lighting scenarios, and RSO configurations. These synthetic copies provide comprehensive training data without requiring equivalent real-world data collection, saving computational resources while maintaining training accuracy.
Data Source
AI summary
A method of generating a simulated image for training a model to detect resident space objects (RSOs), a method of training the model, and a method of detecting RSOs are provided. The method includes generating a foreground of the simulated image including RSOs by calculating coordinates of an imaging satellite at a given time, determining an area of interest given a field of view (FOV) of the imaging satellite, and calculating coordinates of all RSOs and saving the coordinates if the RSO is within the area of interest. The method further includes generating a background of the simulated image including stars by querying a star catalogue database for coordinates of all stars that fall inside the FOV, adding noise, and translating RSO and star coordinates to pixel coordinates. The method further includes generating the simulated image by choosing an imaging mode, setting an exposure time, and rendering the simulated image.