There is a growing need for remote underwater sensing and imaging without the burden of cabled connectivity to the outside world, and the best method for achieving this is with acoustic communications. However, the acoustic link is only one aspect of the problem, as the sensor system needs to interface with the modem, and perhaps with intermediary subsystems to provide some level of data pre-processing. We are developing an integrated processor/modem capability to provide the sensor developer with much of the signal processing capability needed to prepare and store sensor-level data, and to transmit information and receive commands as required. Our prototype development for such a system is focused on still-frame image compression and transmission in support of robotic crawlers in a mine-countermeasure (MCM) mission. Image compression is based on novel wavelet-based algorithms developed in part by the University of California at San Diego (UCSD). The robotic crawlers are a development of Foster-Miller Inc. and the Naval Coastal Systems Station. Benthos provides the acoustic and RF links for command and control and image/sensor transmission. The sensor computer which performs the image compression is a micro-PC with extensive capabilities for R and D applications, including sensor-level data processing, in-situ data fusion, and connectivity to the modem. The modem currently supports signaling up to 2400 bps (raw data rate), although 100 bps to 800 bps are more likely to be used in other than vertical channels. We show examples of full and compressed images, the latter as received via an acomms-to-RF modem gateway buoy. Enhancements expected include incorporation of much higher data rate signaling methods, and customized signal processing for raw sensor-level data.