{HiDyn} Deliverable D5 - Report describing the implementation of an open source HDR algorithm for the generation of traceable luminance images, scaled to one or a few traceable spot measurements of the scene.

This report describes the implementation of an open-source high dynamic range (HDR) algorithm code for the generation of traceable HDR luminance images. The toolbox implements four merging algorithms compatible with uncertainty propagation and provides a complete processing chain including dark correction, raw file handling, and camera calibration.This document aims that anyone in lighting or measurement fields can produce HDR luminance maps using ordinary cameras, helping to standardise measurements for glare, road lighting, and light pollution studies. Imaging devices, such as cameras and Imaging Luminance Measurement Devices (ILMDs), are increasingly used for optical measurements in applications like glare analysis and obtrusive light evaluation. However, the dynamic range of digital imaging sensors is often insufficient for typical scenes in these applications, requiring high dynamic range (HDR) imaging methods.Digital HDR imaging techniques, including merging algorithms, were established in the early 1990s [1]. Today, HDR is commonplace in devices like digital single-lens reflex (DSLR) cameras and smartphones, sometimes employing AI-based algorithms optimized for visually pleasing results, i.e. colour balancing and local tone mapping. In quantitative imaging measurement, rather than aesthetic images, the focus is on producing accurate, SI-traceable luminance maps with well-defined measurement uncertainties, which is the objective of our project.HDR imaging involves capturing multiple low dynamic range (LDR) images at different exposures (e.g., integration times) and combining them using an HDR merging algorithm to create an image with an increased dynamic range. Regardless of the method used to vary exposure, the merging principle remains the same. When a pixel is well-exposed in multiple LDR images, each provides a good estimate of the luminance, whose uncertainty is not limited by noise. The main challenge for HDR algorithms is to combine these images to minimize error, not only due to noise but also from potential systematic errors.In the framework of the joint research project 21NRM01 HiDyn, a review of HDR merging algorithms compatible with uncertainty propagation was conducted. From this review resulted a selection of 4 algorithms for HDR merging that have been implemented in a MATLAB code.