Star Sensor Noise Filtering for Earth Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current star sensing systems, particularly those designed for earth imaging, face challenges in detecting dim stars due to low signal-to-noise ratios, false alarm rates, and reduced accuracy caused by noise spikes, background gradients, and scattered light, which limits their ability to determine accurate line-of-sight angles.
Innovation Solution
The system employs a combination of median and averaging filters to reduce noise and background gradients, along with a threshold module and centroid calculator to enhance star detection and location accuracy, using a 1×3 median filter in the along-scan direction and a 1×12 averaging filter in the along-scan direction, and additional filters in the cross-scan direction to improve signal processing and centroiding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard star detection algorithms are used with earth imaging sensors, then the system can detect stars, but the detection accuracy is poor due to low signal-to-noise ratios and high false alarm rates
Solution Approach 1:
The patent applies preliminary filtering actions before star detection. A median filter is applied first to remove noise spikes, followed by an averaging filter to reduce background gradients. This preliminary noise reduction enables more accurate star detection by improving the signal-to-noise ratio before the actual detection algorithm processes the image data.
Solution Approach 2:
The patent introduces intermediary filtering operations between the raw image data and the star detection algorithm. The median filter and averaging filter act as intermediaries that condition the input data, removing harmful noise components while preserving star signals. This intermediary processing step mediates between the noisy earth imaging data and the detection algorithm, improving both accuracy and reliability.
2Quantity of substance
If the detection threshold is lowered to detect dim stars, then more stars can be detected, but false alarm rates increase due to noise spikes
Solution Approach 1:
The median filter is applied as a preliminary action before thresholding to remove noise spikes that would otherwise trigger false alarms. By eliminating these spurious signals in advance, the system can safely lower the detection threshold to capture dim stars without proportionally increasing false alarm rates.
Solution Approach 2:
The filtering operations serve as intermediaries between the raw signal and the threshold comparison. They condition the signal by removing noise components, allowing the threshold to be set at a level that is sensitive enough to detect dim stars while remaining robust against noise-induced false alarms.
3Measurement precision
If filtering operations are applied to reduce noise, then detection accuracy improves, but processing time increases
Solution Approach 1:
The filtering process is segmented into two distinct stages: first a median filter to remove noise spikes, then an averaging filter to reduce background gradients. This segmentation allows each filter to be optimized for its specific purpose and enables parallel processing possibilities, reducing overall processing time while maintaining accuracy improvements.
Solution Approach 2:
The patent applies filtering operations selectively - the median filter with a small kernel size to remove only the most harmful noise spikes, and the averaging filter with an optimized window size to reduce background gradients to an acceptable level. This partial action approach achieves sufficient noise reduction for accurate star detection without applying excessive filtering that would unnecessarily increase processing time.
Data Source
AI summary
A star sensor includes (a) a scan mirror for scanning at least one star; (b) a detector array, coupled to the scan mirror, for detecting the one star; and (c) a processor, coupled to the detector array. The processor includes a first filter configured to reduce noise spikes in the detected one star, and provide a detection mask of filtered data. Also included is a second filter configured to reduce non-contiguous samples in the detection mask. A centroid calculator is included to determine a location of the one star, after the first and second filtering. The first filter includes a median filter, followed by an averaging filter, both configured to filter the one star in an along-scan direction of the scan mirror. The first filter includes another median filter, which is configured to filter the detected one star in the cross-scan direction of the scan mirror. An adder is included to subtract (a) output data from the other median filter from (b) output data from the averaging filter and provide filtered star data to the second filter.