Propeller scarring of seagrass is a common occurrence in shallow coastal systems and is a challenge for managers to quantitatively assess at scale. Multiple platforms exist to acquire imagery, but detecting and classifying scarring in images can be tedious and often resource prohibitive.  The 455,000-acre Nature Coast Aquatic Preserve along Florida’s west coast, is one such system.  To facilitate managers’ needs to map and quantify propeller scarring in this region, three sources of aerial imagery: drone, manned aircraft, and satellite were collected, and two detection and classification approaches: manual tracing and a computer vision/AI model were evaluated.  At scale, drone imagery provided the highest resolution and the most detailed scar detection, but limited windows of favorable flight conditions made full-scale coverage impractical. Manned-aircraft imagery offered broad spatial coverage with good resolution but would be cost-prohibitive for frequent monitoring. Higher-resolution satellite imagery (0.3–0.5 m/pixel) showed strong potential as the most economical option for routine assessments. In the past, detection and classification of imagery was conducted manually, which is time consuming, somewhat subjective, and if multiple analysts are involved increases variability in scar identification.  To increase objectivity and streamline the process, an AI model was trained and applied to the three image sources and compared to manual classification. Model results were often equal to and in some instances, superior to, manual classification. In addition, model output is probabilistic at the pixel scale allowing the user to apply a level of confidence to the model output, whereas manual classification is binomial (scar or no scar).  The model is still being refined but could dramatically shorten the time and cost to post process imagery for the detection of propeller scarring, significantly improving scar mapping and monitoring for managers.