Bird line conflict is closely connected with power grid lines and bird activity habits, and its prevention and control work need to fully consider the construction of power system and bird protection. In this paper, based on the characteristics of bird line conflict research language, the conflict relationship between bird behavior and power line construction as the core content of its prevention and control knowledge map, using the maximum entropy model to train classifiers, to assist in entity relationship extraction under the perspective of the classifier. The knowledge representation learning model is used to complement the knowledge elements in the knowledge map, and the TransE, TransH and TransD models are selected to enhance the entity and relationship representation learning of the knowledge representation learning model from the spatial transformation level, and the ProjE and PTransE models are used to enhance the relationship feature representation ability of the knowledge representation learning model in the form of combining relationship vectors. Through entity-relationship extraction and knowledge representation complementation, this paper constructs a knowledge graph containing 7023 entity nodes, 4526 attribute nodes, and 7381 entity-attribute relationships for birdline conflict prevention and control. The knowledge graph can clearly show the types of bird line conflicts, bird species related to the conflicts, the contents of conflict prevention measures and the correspondence between them, which provides a reliable path reference for bird line conflict prevention.
