After several decades of stock decline, the Atlantic sea scallop (Placopecten magellanicus) resource has rebounded to become one of the most valuable fisheries in the United States. The continued sustainability of this fishery is supported by catch limits determined by annual stock projection models. Incidental mortality is an important term in these projection models, but is historically difficult to measure. Current estimates are derived from field experiments that relied heavily on qualitative observations and as a result are based on limited data with low precision. To better quantify incidental mortality, a multiple before—after-control—impact experimental design was used to measure the effect of scallop dredging on the disposition of sea scallops that remain uncaptured on the seafloor following dredging. An autonomous underwater vehicle was used to collect color photographs and side-scan sonar images of the seafloor before and after controlled dredge treatments in the mid-Atlantic and Georges Bank regions. Approximately 170,000 photographs were annotated for instances of mortality. Dredge-induced incidental mortality of 2.5% and 8% was estimated for the mid-Atlantic Bight and Georges Bank sites, respectively, a difference that is likely attributable to the relatively harder substrate of the scallop habitat on Georges Bank that results in greater physical trauma to the uncaptured scallops. This study provides a quantitative estimate of incidental mortality using a noninvasive platform that offers precise mission navigation for repeated surveys. The spatial scale and distribution of the study sites are broad relative to past incidental mortality studies, and the substrate types at each are common in the sea scallop fishery. Estimated mortality rates are lower than the values currently used in fishery stock models and suggest the existing values are conservative, but likely appropriate estimates for management purposes.