Invariant Spectral Markers for Thermal 3D Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Thermal markers are not consistently unique across different spectral ranges, leading to challenges in high-quality 3D reconstruction and alignment in thermal imagery due to variable temperatures and lighting conditions, which affects photogrammetry and reconstruction processes.
Innovation Solution
Spectral markers with invariant spectral signatures, designed to maintain consistent spectral values through segmented areas with contrasting patterns, allowing for reliable detection and alignment across multiple images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If thermal markers are used for alignment and 3D reconstruction, then alignment accuracy can be improved, but the markers are not consistently unique across different spectral ranges and their appearance varies with temperature changes
Solution Approach 1:
The spectral marker is divided into multiple segmented areas with contrasting patterns (e.g., hot and cold regions arranged in specific geometries). This segmentation creates distinct spectral signatures that remain identifiable across different spectral ranges and temperature conditions, resolving the reliability issue while maintaining alignment precision.
Solution Approach 2:
The marker uses materials with contrasting spectral emissivity properties (appearing as different 'colors' in thermal imagery) that maintain their distinguishability across varying temperatures. The contrasting patterns ensure the marker remains identifiable in both visible and thermal spectral ranges, addressing the consistency problem.
2Ease of manufacture
If visible spectrum markers (such as small white balls, polka-dots, QR codes) are used for photogrammetry, then alignment can be achieved, but these markers may appear in different colors or lose contrast when sensed using IR sensors
Solution Approach 1:
The spectral marker is designed to function across multiple spectral ranges (visible and thermal infrared) simultaneously. By incorporating materials with contrasting emissivity and geometric patterns, the marker maintains detectability and distinguishability in both visible photography and thermal imaging, achieving universal applicability across different sensing modalities.
Solution Approach 2:
The marker uses composite material structures with different thermal emissivity properties arranged in contrasting patterns. This composite approach ensures the marker appears distinct in both visible light (through geometric patterns) and thermal radiation (through emissivity contrasts), resolving the adaptability issue across spectral ranges.
3Productivity
If standard photogrammetry pipelines are used for thermal images, then 3D reconstruction can be performed, but the results are poor due to variable lighting conditions and lack of consistent reference points
Solution Approach 1:
Spectral markers are pre-placed in the environment before thermal image capture to provide consistent reference points. These markers establish known geometric relationships and spectral signatures in advance, enabling the photogrammetry pipeline to achieve high-quality 3D reconstruction by reliably matching features across multiple thermal images, thus improving both accuracy and efficiency.
Data Source
AI summary
A method for producing multi-dimensional spectral models of a production system. Input data is received from a plurality of spectral markers that are disposed in the production system, the spectral markers providing reference points for multiple readings taken by mobile or stationary sensors to be matched. Each of the spectral markers within the three-dimensional space are read to determine a unique spectral signature corresponding to each one of the spectral markers, the spectral signature having a pattern of spectral values that is unique to each one of the corresponding spectral markers. The determined spectral signatures of each of the plurality of spectral markers are associated with a unique identification which are then provided to a robot within the three-dimensional space. The multi-dimensional spectral model is then reconstructed using the assigned locations. The spectral markers also double as sensors and can transmit readings along with their unique spectral signatures.