Scene Image Retrieval Based on Salient Local Feature Aggregation and Geographic Information

In order to improve the mean average precision (mAP) of scene image retrieval, this paper proposes a scene image retrieval algorithm based on salient local feature aggregation and geographic information, and establishes a scene database that can be verified by the algorithm. On the basis of “probabilistic proposal” selection and weighted aggregation of local detectors of PWA algorithm, image retrieval is carried out by means of geographic information filtering. The addition of geographic information provides more reference features for scene image retrieval, which can improve the accuracy and efficiency of the algorithm. The construction of scene image database with geographic information and scalability provides a test basis for similar algorithms