**6. Conclusions**

Regional geochemical data of 1:200,000 in the northern Daxinganling area were rasterized using a method that triangulates a planar set of points. Consequently, a multilayered image database containing 39 elements/oxides was formed. The images were enhanced using an image enhancement technique and algebraic operations. The images were handled as multidimensional vector data. Accordingly, hyperspectral tools could be used for the processing. The geochemical signatures of deposits were extracted from the images. Enriched and depleted elements were distinguished by comparing them with regional geochemical statistics. The geochemical signatures represented the geochemical characteristics of ore deposits. The rock types were classified using the K-Means method, which assisted in the regional geological mapping, especially in the areas of dense forest. The geochemical signature of a typical ore deposit was processed by SAM, which determined the similarity between the deposit and pixels in the region. The prospecting target area was determined according to the angle. With increased geochemical data sampling density, as well as further integration with other geophysical, geological, and remote sensing data, rasterized geochemical images can be fully used in the future.
