Evaluating Roadway Lighting Conditions with Advanced Image Processing Techniques

Abstract: Street lighting is a significant factor in road safety. Traditionally, the assessment of the adequacy and distribution of the street lighting over the large areas has been labor-intensive, often by human measurement using simple light meters. This paper presents a new image processing approach that can effectively assess the comparative lighting conditions of the road surfaces and streets using video image processing techniques. The study areas are two major interchanges in Louisville, Kentucky: I-64 at 22nd Street and I-264 at Bank Street, both of which recently had new solar-powered streetlights installed. Solar-powered lights are sustainable and do not require wired connectivity to the power grid to install. As this is the first use of solar-powered lights for roadway lighting in Louisville, the objective of this project is to assess their effectiveness using video image processing. For this, a GPS-enabled, vehicle-based light intensity imagery collection system was used that is capable of data collection in drives through the study areas. This system captures continuous footage of the road surface that can then be processed by use of more advanced image-processing techniques, including grayscale conversion, region of interest (ROI) extraction, pixel intensity calculation, colormap application, time and GPS overlays, and frame-by-frame processing. These techniques enable the transformation of video data into grayscale images, allowing for precise pixel intensity analysis to estimate brightness levels in the specified ROIs. This study assesses comparative lighting conditions using both visualization techniques and statistical analysis. Lighting is depicted in the form of heatmaps, time-series, descriptive statistics, statistical tests, and color mapped videos. The video-based technique increases resolution in the lighting assessment and also provides a feasible and inexpensive method, relative to existing methods, of determining the lighting conditions over large areas.