Histogram Adjustment of Images for Improving Photogrammetric Reconstruction

Area of Science:

• Photogrammetry and Computer Vision
• Image Processing

Background:

• Photogrammetric reconstruction accuracy is highly dependent on input photograph quality and acquisition conditions.
• Enhancing raster image quality is crucial for improving photogrammetric outcomes.

Purpose of the Study:

• To propose and evaluate methods for improving raster images to enhance photogrammetric reconstruction accuracy.
• To investigate the impact of color histogram modifications using RGB and CIE L*a*b* color models.

Main Methods:

• Modification of color image histograms.
• Selection of optimal color channels (RGB and CIE L*a*b*) for image enhancement.
• Development of a quality assessment methodology using reference models and positional statistics.

Main Results:

• Proposed image enhancement methods significantly improve photogrammetric reconstruction quality.
• Methods utilizing the luminance channel of the CIE L*a*b* color model demonstrated superior performance.
• Histogram equalization (HE) showed high efficiency, though results varied across different object types and tests.

Conclusions:

• Image enhancement techniques, particularly histogram modification, are effective in boosting photogrammetric reconstruction accuracy.
• The luminance channel of the CIE L*a*b* color model offers significant advantages for photogrammetric applications.
• Further research may refine HE and other methods for consistent improvements across diverse datasets.