Efficient and reliable methods of canal inspection are necessary for supporting the maintenance decision-making. One focus of such decision-making is predicting and preventing water losses in a canal system by identifying cracking sections of concrete canal facilities. Current practice requires engineers to spend much time and money on a visual inspection and engineering surveying for identifying sections of canals that are subject to water losses. Such manual efforts could hardly achieve detailed assessment of underwater conditions of canals. Emerging imaging technologies, such as terrestrial and mobile laser scanning can collect high-resolution spatial data for supporting detailed condition assessment of canal facilities. Unfortunately, terrestrial and underwater imaging systems are expensive and tedious to use for mapping a large canal system, considering both the time and costs for high-resolution imagery data collection and processing. For example, maintaining Salt River Project (SRP)'s 131-mi canal system with mobile and terrestrial laser scanning can cost thousands of dollars every day, and even more time and money for hiring professionals to process terabytes of 3D imageries. This research examines an integrated analysis method of aerial and terrestrial imagery data to reduce the time and costs for mapping water losses of a canal system without compromising the comprehensiveness of the inspection. The method first applies remote sensing techniques to process satellite images, which are much cheaper than terrestrial and underwater laser scanning, to extract typical environmental features along canals, including land surface temperature (LST), humidity, and plant cover index. The algorithm then analyzes spatiotemporal changes of those features along the canals and identifies sections of canals that are more likely to have water losses (e.g., sections that have a lower temperature and higher humidity than adjacent sections). Terrestrial and underwater laser scanning then can get detailed data to inspect those anomalous canal sections. A case study shows the potential of the proposed two-stage approach that uses both satellite and terrestrial imageries for efficient and effective water loss mapping of a large canal system.