A problem is addressed of minimization of the number of measurements needed for image acquisition and reconstruction with a given accuracy. In last several years, the compressed sensing approach to solving this problem was advanced, which promises reducing the number of required measurements by means of obtaining sparse approximations of images. However, the number of measurements required by compressive sensing substantially exceeds the theoretical minimum defined by sparsity of the image sparse approximation. In the paper, a sampling theory based method of image sampling is suggested that represents a practical and substantially more economical alternative to the compressed sensing approach. Presented and discussed are also results of experimental verification of the method, its possible applicability extensions and some its limitations.