To address the problem that traditional interpolation methods are difficult to effectively fuse rain gauge and microwave link rainfall observation data, a collaborative reconstruction framework for rainfall fields based on a stochastic hybrid Whittaker-Shannon algorithm was proposed. By constructing a synthetic rainfall reference field, the two-dimensional precipitation reconstruction performance of single-source interpolation of rain gauges and dual-source fusion of rain gauges and microwave data was systematically compared. At the same time, an image smoothing technique was used to mitigate the problems of excessive noise and lack of smoothness of the stochastic reconstruction method in rainfall field construction. The results show that the spatial reconstruction accuracy of the stochastic hybrid Whittaker-Shannon algorithm is significantly positively correlated with the rain gauge density and microwave link coverage. Moreover, as the number of possible reconstructed fields increases, the accuracy is better; however, when the number of possible reconstructed fields exceeds 20, the overestimation rate of rainfall extent increases to 80%. The 5×5 mean filter can reduce the root mean square error by 60%, which is significantly superior to the median filter and other window scale schemes. Microwave link data have complementary value in enhancing the spatial resolution (up to the 1 km scale) and dynamic representation capability of rainfall fields, providing a scalable and highly promising technical pathway for real-time data assimilation of urban hydrological models under the 5G communication context.