The presence and nature of uranium minerals subject to high leachability is one of the most significant factors in the exploitation of low-grade ores. The detection of these minerals is a challenging task due to their small size and low abundance. This paper presents a novel method of data processing that allows distinguishing and visualization of micron-scale U minerals. For this purpose, core samples of U-bearing sandstones from the Brevniste deposit (Czech Republic) were used. After their study by inductively coupled plasma spectroscopy, optical and scanning electron microscopy (SEM-EDS) and X-ray diffraction analysis, the presence of authigenic grains of U minerals occurring in various associations and their chemical compositions were confirmed by an electron microprobe. However, the use of these conventional instruments for the analysis of heterogeneous ore material with valuable information hidden, showed the necessity for a special data treatment. Matrix transformation of raw SEM-EDS data allowed for even more accurate visualizations such as contour and point maps of elemental distribution. Then, mathematical and visual correlations of transformed data revealed relationships among some measured elements (Ca, P, U, Zr, Nb, Fe, S) and their spectral overlaps. Conditions of ore formation were predicted based on the visualizations, correlations and the paragenesis of ningyoite in uranium ore. Herein suggested approach can help to identify economically significant minerals in complex mineralizations worldwide and increase the mining potential of the deposits.