In recent years, applications using wireless technology has dramatically increased. This has led to the challenge of improving the spatio-temporal spectrum utilization. Dynamic spectrum access is a method of opportunistically accessing the spectrum. Identifying the unused spectrum is one of the key challenge for dynamic spectrum access. Regulatory bodies recommend the usage of centralized database to maintain the unused spectrum at various geographic locations. These databases are constructed with the help of propagation models. Propagation models are not accurate enough for effcient spectrum utilization. To encounter this problem, we propose a cloud-based and crowd- augmented spectrum mapping database by combining the manual spectrum measurements from various mobile devices. Then, we show how optimal sampling can help in achieving high resolution, high accurate spectrum mapping in minimal cost.