Indoor Place Recognition System for Localization of Mobile Robots

Sahdev, Raghavender; Tsotsos, John K.

doi:10.1109/crv.2016.38

Cited by 22 publications

(26 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Since most shallow learning algorithms perform poorly on raw image data, both global and local image descriptors have been used to extract image features [23,24,25,26,27]. These feature descriptors extract textural features such as edges or bright and dark spots, which can be used to categorize the image [15]. Due to the high variability of environments conditions, handcrafted descriptors struggle with robustness [1,17].…”

Section: Related Workmentioning

confidence: 99%

“…We trained and evaluated our recognition module on the York University 11 and 17 places and the Rzeszów University 16 places datasets [15,19]. The first dataset consists of 11 indoor places captured by two robots (Pioneer and Virtual Me) under different lighting conditions (daytime and nighttime) at York University.…”

Section: Evaluation Data and Data Pre-processingmentioning

confidence: 99%

“…Robot localization and navigation has been extensively studied, using Lidar [11], Sonar [12], GPS [13] and sensor information [14]. Place recognition on the other hand is mostly based on categorizing monocular images into predefined classes [15]. Robust place recognition is one of the most difficult challenges, due to the high complexity of the places themselves [16], changing lighting and viewpoint condition [15], occlusions of objects and landmarks [16] and the presence of other dynamic objects [17].…”

Section: Introductionmentioning

confidence: 99%

“…We use the MobileNetV2 [18] architecture -a highly optimized convolutional neural networks developed for mobile and embedded devices -to build our place recognition module. We train and evaluate our module using the York University 11 and 17 places [15] and the Rzeszów University 16 places dataset [19], outperforming current benchmarks. Since both datasets are relatively small we use a transfer learning approach based on the explicit inductive bias L 2 -SP [20].…”

Section: Introductionmentioning

confidence: 99%

“…Our approach is robust against severe changes in lighting and viewpoint conditions [1] and shows good performance in different environments [15].…”

mentioning

confidence: 96%

See 4 more Smart Citations

Development of a Highly Precise Place Recognition Module for Effective Human-robot Interactions in Changing Lighting and Viewpoint Conditions

Baumgartl¹,

Buettner²

2020

Proceedings of the Annual Hawaii International Conference on System Sciences

View full text Add to dashboard Cite

We present a highly precise and robust module for indoor place recognition, extending the work by Lemaignan et al. and Robert Jr. by giving the robot the ability to recognize its environment context. We developed a full end-to-end convolutional neural network architecture, using a pre-trained deep convolutional neural network and the explicit inductive bias transfer learning strategy. Experimental results based on the York University and Rzeszów University dataset show excellent performance values (over 94.75 and 97.95 percent accuracy) and a high level of robustness over changes in camera viewpoint and lighting conditions, outperforming current benchmarks. Furthermore, our architecture is 82.46 percent smaller than the current benchmark, making our module suitable for embedding into mobile robots and easily adoptable to other datasets without the need for heavy adjustments.

show abstract