DETECTION OF FLUORESCENT FLOATING-BEAD TRACERS USING AN IMAGE ANALYSIS ALGORITHM

In a previous study by our group, fluorescent hydrogel tracer beads were developed as proxies for research into the fate and transport of light non-aqueous phase liquids (LNAPLs) in karst aquifers. The beads are made of calcium alginate, with additives to adjust their density and pigments to aid in detection. The major challenge in the previous study was quantification of bead transport in water. In this study, we developed and tested an MATLAB^® algorithm that analyzes a video of bead movement using built-in functions. Beads were released in flowing water, a low-power ultraviolet (UV) LED light source was used to excite the pigments, and fluorescence was recorded on a video camera. The algorithm takes the video recording and splits it into frames. Then each frame is analyzed for average pixel intensity. The videos used to test the algorithm were recorded in dark conditions with a low-cost digital camera at an acquisition rate of 30 frames per second. Reflection of UV light at the water surface was filtered by isolating the green and red channels from the RGB frames in the algorithm. Preliminary results show that frames with fluorescent beads have average pixel intensity proportional to the number of beads in the frame, while frames with no beads have average pixel intensity equal to zero. The preliminary findings suggest that the algorithm developed in this study can be used to easily detect the transit of fluorescent hydrogel beads. Future work will focus on using a camera with a higher acquisition rate to minimize distortion, testing of limited-wavelength filters, and optimization of LED-pigment combinations to maximize the signal.