Sorting gangue from raw coal is an essential concern in coal mining engineering. Prior to separation, the location and shape of the gangue should be extracted from the raw coal image. Several approaches regarding automatic detection of gangue have been proposed to date; however, none of them is satisfying. Therefore, this paper aims to conduct gangue segmentation using a U-shape fully convolutional neural network (U-Net). The proposed network is trained to segment gangue from raw coal images collected under complex environmental conditions. The probability map outputted by the network was used to obtain the location and shape information of gangue. The proposed solution was trained on a dataset consisting of 54 shortwave infrared (SWIR) raw coal images collected from Datong Coalfield. The performance of the network was tested with six never seen images, achieving an average area under the receiver operating characteristics (AUROC) value of 0.96. The resulting intersection over union (IoU) was on average equal to 0.86. The results show the potential of using deep learning methods to perform gangue segmentation under various conditions.

A small amount of research has been conducted in dealing with pavement shadow–affected cracks detection. Hence, this research aims to develop the illumination compensation model (ICM) and $k$ -means clustering algorithm–based crack detection considering the influence of pavement shadows. First, the shadow area was divided into the umbra area and the penumbra area according to the illumination mechanism. Then, the shadow removal methods for different areas were analyzed separately. Since the intensity of the umbra shadow area changes homogeneously, the ICM approach can be a convenient way for shadow removal. While the intensity of penumbra area changes drastically, the cubic sample interpolation operation was conducted in advance, followed by ICM to finalize the shadow removal. After that, the $k$ -means clustering algorithm was used to extract the crack region from the road background. Finally, based on the segmented binary crack image, the orientation, crack length, width, aspect ratio, area, and blocks were calculated for comprehensive crack-type classification and severity evaluation. Experiments were conducted to compare the performance of the proposed approach with traditional threshold segmentation, Poisson equation, contourlet transformation, and CrackTree, which demonstrated optimistic performance of the proposed method in terms of average precision (93.58%), recall (94.15%), and F-measure (93.86%).