Balancing image quality and energy consumption in visual sensor networks

Applications of wireless sensor networks have attracted a lot of attention recently. Cameras are installed in various locations of a wide area to capture images of targeted objects. Because of constraints in computational capability in these distributed cameras, it may not be feasible to analyze these images in the sensors but they have to be transmitted to a centralized server hop by hop through the sensor network. To reduce the energy used in transmission, the size of the images should be kept small by applying a large compression ratio, which may degrade image quality. This paper studies the tradeoff between image quality and energy consumption. We study the scenario that a number of camera-equipped sensors are taking pictures of the same object, and the pictures of adjacent cameras may overlap. We demonstrate that by allowing intermediate sensors to process the images and combine the overlapping portions, the total energy spent on transmission is reduced subject to a certain degradation in image quality. The tradeoff between image quality and energy consumption of different routing presents an important study on the practicability of visual sensor networks.