Environment monitoring help us to learn the natural environment we live in. The complexity of the many natural phenomena makes it a very difficult task to learn these phenomena. Different sensors assist us in this task from different perspective. In general, to reveal the phenomena in detail requires a large amount of sensors and is often prohibitively expensive. Through our study, we proved that by combining measurements from different types of sensors, we can reduce total cost a great deal while keeping the same fidelity level as exhaustive sensing with one type of sensors.

We developed an algorithm that combine measurements from two types of sensors. From the high level information provided by one type of data, we built field models, which was applied to low level data for field reconstruction. Experimental results demonstrated the algorithm worked effectively.