Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence
2. ENGINEERING AND TECHNOLOGY
Place recognition in a visual SLAM system helps build and maintain a map from multiple traversals of the same environment while closing loops to correct drift accumulated over time. Despite the marked success in visual place recognition research over the past decade, it remains a challenging problem in the context of variations caused due to different times of the day, weather, lighting and seasons. In this paper, we address this problem by progressively training convolutional neural networks in a siamese fashion to generate embeddings that encode semantic and visual features for sequence-aligned image pairs taken at different timescales and viewpoints. We present early results of the approach using Freiburg visual place recognition benchmark dataset consisting of aligned outdoor image sequences taken over extended time periods that include the variations mentioned above.
Ramachandran, S. & McDonald, J. (2019). Place recognition in challenging conditions. IMVIP 2019: Irish Machine Vision & Image Processing, Technological University Dublin, Dublin, Ireland, August 28-30. doi: 10.21427/seen-pq19