An Improved Indoor Navigation Scheme Based on Vision-Language Localization

This paper proposes an improved indoor localization and navigation method based on image matching and text extraction. Compared with traditional indoor localization methods based on WIFI, Bluetooth or laser, the improved method is mainly realized by image matching in computer vision and text extraction in natural language processing. The improved bidirectional A* algorithm is adopted to realize the optimal path planning, and the specific location of indoor localization and path planning diagrams are displayed through an app. Experimental results show that this method which can realize indoor location determination in real scenes, is convenient to use, and has low cost. It can overcome the problems of high hardware requirements, high cost, and poor transferability in the traditional indoor navigation field in terms of localization.