Image Enhancement and Accurate Recognition of Coal and Gangue Based on 3-D Depth Information Guidance

To minimize the impact of external factors on the identification of coal gangue and ensure accurate identification under complex conditions, an analysis of various external influences was conducted. A new image enhancement algorithm leveraging multi-dimensional information fusion and a novel set of features were introduced. The image quality serves as the control coefficient for performing adaptive gray mapping on RGB images, and the mapped RGB images undergo multi-level decomposition at various scales to separate high-frequency and low-frequency components. Depth texture images and regional fading masks are derived from the image depth information. Guided by the regional fading mask, the RGB image is processed at the pixel level in both spatial and frequency domains to enhance authentic textures and suppress artificial textures. By integrating the proposed depth texture features with enhanced gray, texture, and gradient features, accurate identification of coal gangue under complex conditions is achieved. Comparative analysis of recognition results from different algorithms reveals the superior recognition performance and adaptability of the proposed method, with accuracy and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">F</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> scores under four conditions being (97.143%, 0.970), (94.028%, 0.938), (90.310%, 0.902), and (92.237%, 0.920).