Hyperspectral data set of stone cultural relics in high-precision machine vision scene.

In the contemporary era, which is characterized by its emphasis on intelligence and technological advancement, the rapid evolution of machine vision technology has significantly enhanced the efficiency of artificial systems and yielded noteworthy achievements across numerous domains. Nevertheless, in certain specialized fields, such as the high-precision identification of the deterioration modes of stone cultural relics, the limitations of this technology persist. The primary rationale for this limitation is the lack of multi-dimensional and high-precision data support. In response, our team has collaboratively developed hyperspectral acquisition equipment with professional manufacturers and successfully established the world's first hyperspectral dataset of stone cultural relics (HD-SC-1). This dataset encompasses 28,193 hyperspectral layers from China's Dazu Rock Carvings. The spectral range of the dataset spans from 400 to 1000 nm, with an image resolution of 2048 × 2046 pixels. The dataset expands the scope of data on stone cultural relic deterioration and enriches the data resources for the protection of stone cultural relics, thereby revitalizing the conservation and restoration of world cultural heritage.