A capsule is a group of neurons, whose activity vector represents the instantiation parameters of a specific type of entity. In this paper, we explore the capsule networks used for relation extraction in a multi-instance multilabel learning framework and propose a novel neural approach based on capsule networks with attention mechanisms. We evaluate our method with different benchmarks, and it is demonstrated that our method improves the precision of the predicted relations. Particularly, we show that capsule networks improve multiple entity pairs relation extraction 1 . * Corresponding author. 1 In this paper, multiple entity pairs relation extraction refers to multiple entity pairs in a single sentence and each pair of entities contains only one relation label. Related WorkNeural Relation Extraction: In the recent years, NN models have shown superior performance over approaches using hand-crafted features in various tasks. CNN is the first one of the deep learning models that have been applied to relation extrac-

The ground cracks in the goaf will affect the safety of the local people and the ecological environment. Therefore, the three-dimensional model of the ground cracks can provide certain decision support for the coal mine safety production managers. In this paper, a three-dimensional model method for ground cracks based on images acquired by drones is proposed. Firstly, based on a small amount of original borehole data, Kriging interpolation method is used to generate enough stratum information, and the grid model is used to establish the original stratum three-dimensional model. Secondly, the image acquired by the drone is needs to be preprocessed, and the corner data is obtained as the key point of the crack by using the Harris algorithm. Thirdly, generate the three-dimensional model of the ground crack. It is inserted into the original three-dimensional model of the original stratum, and finally a three-dimensional model of the stratum containing the ground crack is obtained. The experimental verification shows that the method can effectively simulate the ground crack.