Learning Model for SAR Signal Meta-Information Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
SAR image generation processes that filter received signals to reduce data volume lead to missing information and false detections, affecting the precision and accuracy of object and environmental observations.
Innovation Solution
A learning model and signal processor are used to process received signals from a flying object, converting them into meta-information with high precision and accuracy, enabling accurate observation and change detection by associating the signals with meta-data using machine learning techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If filtering is applied to the received signal to reduce data volume, then the computational burden of compression is reduced, but information may be lost or falsely detected
Solution Approach 1:
The patent applies filtering as a preliminary action before compression processing to reduce data volume. By pre-filtering the received signal in the frequency domain, the system removes unnecessary frequency components before the computationally intensive compression process, thereby reducing computational burden while preserving essential information for accurate observation.
Solution Approach 2:
The patent changes the parameter of the received signal by applying frequency domain filtering. This parameter change selectively removes certain frequency components while preserving others, optimizing the balance between data reduction and information retention. The filtering operation transforms the signal in the frequency domain, keeping only the most relevant frequency ranges for subsequent compression and observation.
2Quantity of substance
If filtering removes parts of the received signal, then data volume is reduced, but precision of observation is affected
Solution Approach 1:
The patent applies local quality by selectively filtering specific frequency components of the received signal rather than uniformly processing all data. The filtering operation targets particular frequency ranges that are less critical for observation, thereby reducing data volume while preserving the quality and precision of the most important signal components.
Solution Approach 2:
The patent changes parameters of the received signal through frequency domain filtering, transforming the signal to selectively remove or attenuate specific frequency components. This parameter transformation reduces the quantity of data while maintaining the precision of observation by preserving the most informative frequency ranges.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The purpose of the present invention is to provide a learning model, a signal processor, a flying object, and a program that enable appropriate observation of the situation of an observed object or the environment around the observed object. The learning model is learned by using teaching data with a first received signal as input, the first received signal being based on a reflected electromagnetic wave that is an electromagnetic wave emitted to a first target area and then reflected, and with first meta-information as output, the first meta-information corresponding to the first received signal and having a predetermined item, so as to input a second received signal based on a reflected electromagnetic wave that is an electromagnetic wave emitted to a second target area and then reflected, and to output second meta-information corresponding to the second received signal and having a predetermined item.