In this work, the focus was on the analysis of resource consumption in a sensor-based, mobile AR application, SARAS. As the resources, it uses motion sensors and the graphics API in addition to GPS, the camera and the web services for the content of bank data. Firstly, CPU profiling using the methods available in Android platforms were examined and improvements suggested by these methods are applied to SARAS. The functions and methods that consume high power were determined and modified. These improvements reduced the resource consumptions up to %35. Afterwards, a detailed energy analysis was performed under different GPS and sensor processsing settings. Finally, the comparison with other applications were performed in terms of energy consumption. As a result, it has been observed that SARAS consumes similar amount of energy compared to similar commercial applications like LAYAR.