Image-based Indoor Positioning: Current Status and Challenges

Abstract. The loss of GNSS (Global Navigation Satellite Systems) signals in enclosed spaces has revealed the need for potential positioning solutions in indoor environments that require high accuracy. Although the solutions developed so far have not yet provided a universally accepted system in terms of metrics such as cost, feasibility, and performance, image-based positioning systems have recently become the focus of researchers' interest in response to the current search. The literature review conducted revealed a gap in the literature, as no study presenting the current developments in this field was found. This study aims to fill this gap and provide researchers with a reference source. Accordingly, studies between 2020 and 2025 were collected using the keywords “vision-based,” “image-based,” “camera-based,” “visual,” “indoor localization,” “indoor positioning,” “indoor navigation,” “indoor tracking,” “visual SLAM,” “VIO”, “simultaneous localization and mapping”, “feature matching”, “image matching”, “feature-based”) were collected. Since 2025 has not yet been completed, a prediction was made using linear regression with data from previous years. The collected publications have been disaggregated using DOI numbers, and studies not containing the keywords (“visual”, “image”, “camera”, “SLAM”, ‘feature’, “indoor positioning”, “indoor navigation”, “indoor tracking”, “indoor localization”) have been excluded. Existing studies were examined as traditional methods and deep learning-based approaches, and deep learning-based approaches were found to be superior to traditional methods in terms of speed, reliability, and accuracy metrics. However, surfaces with low texture, moving environments, and variable lighting conditions remain limitations for both methods. Future studies should test the developed systems in different areas and scenarios.