研究目的
To develop and implement a method for correcting temperature patterns in infrared thermography by using 3D information to account for variations in emissivity and distance due to complex body shapes and constrained camera positions.
研究成果
The proposed method successfully corrects temperature patterns in infrared thermography by incorporating 3D geometric data, reducing standard deviations in temperature distributions and enabling effective damage recognition even under constrained camera positions. Future work includes automating material recognition from point clouds.
研究不足
The method requires accurate 3D point cloud data and may not handle areas without correspondence between the point cloud and thermal image. Effects due to changes in pixel footprint area with incidence angle are not fully addressed, and manual intervention may be needed for image alignment in some cases.
1：Experimental Design and Method Selection:
The study proposes a procedure that integrates thermal imaging with 3D point cloud data to compute pixel-specific emissivity and distance values. It uses MATLAB for implementation, involving steps like image registration, geometric modeling, and temperature computation based on physical laws.
2：Sample Selection and Data Sources:
Data were acquired from Caorle's bell tower, a historic cylindrical masonry structure, using a thermal imager and photogrammetric point cloud. Thermal images were taken from multiple camera positions and poses during night cooling.
3：List of Experimental Equipment and Materials:
FLIR T620 thermal imager (480x640 pixel bolometric array in LWIR band), a camera for photogrammetry (4288x2848 pixel matrix), MATLAB software with Computer Vision Toolbox, ExifTool for metadata handling.
4：Experimental Procedures and Operational Workflow:
The procedure includes generating gray level temperature images, projecting point clouds, image registration, material area selection, temperature computation using geometric information, and de-projection to output corrected temperature patterns.
5：Data Analysis Methods:
Temperature patterns were analyzed statistically (mean, standard deviation) in selected areas to evaluate the effectiveness of the correction method. Damage recognition was assessed using temperature contrast methods.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
