The close-range photogrammetry is a discipline of precise metric measurement of features with applications ranging from the deformation of architectural structures to the medical diagnosis. One of the limitations on the precision of the obtained images is the large differences in depth. On the other hand, obtaining pictures that focus on every direction with lenses used in close range photogrammetry is mechanically impossible. Multiple images focused to different distances are necessary in order to solve such a depth problem. The clear parts of the images are brought together, so that a clear image focused to every direction is obtained. This joint image becomes more suitable for the feature extraction as well as for dividing and classifying processes. In this study, image joining has been conducted with the spatial frequency method. The unclear parts in the images have to be determined so as to be able to bring these images together. The average filtering method has been employed to accurately determine the prescribed clarity level and the gray value levels of the image. For this procedure, software has been developed in C++ programming language that conducts average filtering. The conducted experiments show that the spatial frequency method proves to be a very efficient method to construct a multi-focus image. Due to the precision limitations of the close-range photogrammetry arising from the large differences in the depth, many applications of ultra-high precision measurements such as industrial measurements still heavily rely on the classical surveying methods. The results show that the proposed image-fusion method dramatically reduces the error budget of close-range photogrammetry down to a half and enables much higher precision applications.