| Abstrakt: |
Image geotagging is a process where geographic coordinates are attached to an image. Mobile-based geotagging application has many advantages, viz. real-time monitoring, ensuring data authenticity etc. Since an ordinary mobile camera cannot interpret the geotagged images, they are manually analysed later for a specific purpose. Therefore, the human interpreters are to put their time and effort to analyse the images. This becomes difficult when the number of images is more. The heterogeneity of captured images, in terms of intensity, viewing angle etc., limits the application of traditional image processing techniques for automatic image interpretation. Hence, artificial intelligence (AI)–based image processing technique needs to be employed that enables machines to learn from instances and provide assistance in field photo interpretation. In the present work, a smartphone-based application, embedded with enhanced capabilities of AI and geospatial technology, has been developed using open-source technology. The application employs AI to detect certain categories' semantic objects and automatically generates their details. The mean of detection precision, recall and F1 score are estimated as 0.96, 0.91 and 0.93, respectively. The present work successfully demonstrates the use of open-source technology for AI-enabled geotagging and dissemination of ground information through WebGIS application. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink