The University of Connecticut and the U.S. Geological Survey (USGS) collected low-altitude (30-50 m above ground level) airborne visible-light imagery data via a quadcopter, small unoccupied aircraft system (UAS or ‘drone’) deployed along two tributary confluence locations within the Housatonic River: Mill Brook (latitude: 42°52’18” N, longitude: 73°21’48” W) and Furnace Brook (latitude: 41°49’16” N, longitude: 73°22’17” W). Both tributary confluence sites serve as critical summer thermal refuge for cold water-adapted poikilotherms. The objectives for this data collection included the creation of high-resolution orthomosaic images of the two tributary confluences to infer bank and instream structures and mixing processes at the tributary confluences. Detailed site-scale maps such as these are important tools for managers and researchers aiming to protect and conserve populations at risk. The UAS (Mavic 2 Zoom, DJI Enterprises) was flown several times per day, at wind speeds below 10 mph, capturing RGB imagery from March 24-25, 2021. The UAV flights collected single RGB JPG images at 30-50m above ground level using the double-grid flight pattern on the third-party app Pix4D Capture (https://www.pix4d.com/product/pix4dcapture). The images were stitched automatically into several orthomosaic images using Agisoft Metashape (Agisoft LLC, St. Petersburg, Russia) software as described in the ‘processed_data’ subfolders of this data release. Structure from Motion techniques were also applied to the visual imagery to derive time-specific, digital surface models (DSM) of the exposed banks and some subsurface features.