Hierarchical visual mapping with omnidirectional images

Korrapati, Hemanth; Uzer, Ferit; Mezouar, Youcef

doi:10.1109/iros.2013.6696882

Cited by 6 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this sense, large amounts of data will be essential to test these approaches, where very recent large-scale datasets (Carlevaris-Bianco et al, 2016) could be interesting to perform new experiments. In addition, more datasets based on fish-eye and panoramic images could be also evaluated in future research, such as the IPDS (Korrapati et al, 2013) and Rawseeds (Ceriani et al, 2009) datasets.…”

Section: Results About the Efficiency Of Ablementioning

confidence: 99%

“…On the other hand, some state-ofthe-art algorithms trust in panoramic images for localization across the different seasons of the year (Valgren and Lilienthal, 2010). Additionally, other works also use panoramic views for a more robust loop closure detection (Murillo et al, 2013;Korrapati et al, 2013;Korrapati and Mezouar, 2017). The main advantage of panoramas is that they allow a visual perception of the environment in all the possible orientations, which can be used for detecting places revisited in other direction.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Are you ABLE to perform a life-long visual topological localization?

Arroyo

Alcantarilla²,

Bergasa

et al. 2017

Auton Robot

View full text Add to dashboard Cite

show abstract

Section: Results About the Efficiency Of Ablementioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Are you ABLE to perform a life-long visual topological localization?

Arroyo

Alcantarilla²,

Bergasa

et al. 2017

Auton Robot

View full text Add to dashboard Cite

show abstract

“…Fast Appearance-Based Map (FAB-MAP) [12] is one of the typical applications of the BoVW and has become a standard baseline in loop closure detection algorithms. Fisher Vector (FV) [17,18] and vector of locally aggregated descriptors (VLAD) [19,20] are derived from the basic idea of BoVW and become mainstream algorithms for visual SLAM [21,22]. Fisher vectors utilize a Gaussian Mixture Model to generate codebook and local descriptors are vector-quantized to visual words.…”

Section: Introductionmentioning

confidence: 99%

Loop Closure Detection in RGB-D SLAM by Utilizing Siamese ConvNet Features

Gang

Zhang

et al. 2021

Applied Sciences

View full text Add to dashboard Cite

Loop closure detection is a key challenge in visual simultaneous localization and mapping (SLAM) systems, which has attracted significant research interest in recent years. It entails correctly determining whether a scene has previously been visited by a mobile robot and completely establishing the consistent maps of motion. There are many loop closure detection methods that have been proposed, but most of these algorithms are handcrafted features-based and perform weak robustness to illumination variations. In this paper, we investigate a Siamese Convolutional Neural Network (SCNN) to solve the task of loop closure detection in RGB-D SLAM. Firstly, we use a pre-trained SCNN model to extract features as image descriptors; then, the L2 norm distance is adopted as a similarity metric between descriptors. In terms of the learned features for matching, there are two key issues for discussion: (1) how to define an appropriate loss as supervision (utilizing the cross-entropy loss, the contrastive loss, or the combination of two); and (2) how to combine the appearance information in RGB images and position information in depth images (utilizing early fusion, mid-level fusion or late fusion). We compare our proposed method of different baseline by experiments carried out on two public datasets (New College and NYU), and our performance outperforms the state-of-the-art.

show abstract