StreetMap - Mapping and Localization on Ground Planes using a Downward Facing Camera

Chen, Xu; Vempati, Anurag Sai; Beardsley, Paul

doi:10.1109/iros.2018.8594157

Cited by 16 publications

(59 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…State-of-the-art methods solving the introduced localization task rely on the identification of visual feature correspondences to determine the pose of a query image relative to the mapped reference images [3], [13], [16]. They extract features, i.e.…”

Section: Related Workmentioning

confidence: 99%

“…1 Robert Bosch GmbH, Hildesheim, Germany SchmidJanFabian@gmail.com 2 VSI Lab, CS Dept., Goethe University, Frankfurt am Main, Germany 3 Norwegian Open AI Lab, CS Dept., NTNU Trondheim, Norway…”

Section: Supplementary Materialsmentioning

confidence: 99%

“…This approach has some advantages over approaches using forward-facing cameras: it does not have any issues with occlusions of the surrounding, it works in environments without static landmarks, and it can be made independent of the surrounding lighting. Previous work has shown that reliable high-accuracy pose estimates can be achieved with ground texture based localization on typical ground texture types, such as asphalt and concrete [3], [4], [5], [6], [11], [13], [16].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Model-Based Parameter Optimization for Ground Texture Based Localization Methods

Schmid,

Simon,

Mester

2021

Preprint

View full text Add to dashboard Cite

A promising approach to accurate positioning of robots is ground texture based localization. It is based on the observation that visual features of ground images enable fingerprint-like place recognition. We tackle the issue of efficient parametrization of such methods, deriving a prediction model for localization performance, which requires only a small collection of sample images of an application area. In a first step, we examine whether the model can predict the effects of changing one of the most important parameters of feature-based localization methods: the number of extracted features. We examine two localization methods, and in both cases our evaluation shows that the predictions are sufficiently accurate. Since this model can be used to find suitable values for any parameter, we then present a holistic parameter optimization framework, which finds suitable texture-specific parameter configurations, using only the model to evaluate the considered parameter configurations.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Supplementary Materialsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Model-Based Parameter Optimization for Ground Texture Based Localization Methods

Schmid,

Simon,

Mester

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…There are researches [23]- [26] that demonstrate modeling confidence and uncertainties can improve neural network performance; besides, in both theory and practices, e.g. [27]- [29], ensemble of multiple estimators (in our context, different sensors) can generally improve prediction performance. Our proposed method differs in that we focus the on confidence of sensor inputs and hence provide a novel ensemble method for sensor fusion.…”

Section: Related Workmentioning

confidence: 99%

Brake Maneuver Prediction – An Inference Leveraging RNN Focus on Sensor Confidence

Liu

Koch

Gahr

et al. 2019

2019 IEEE Intelligent Transportation Systems Conference (ITSC)

View full text Add to dashboard Cite

In recent years, driver behavior analysis has led to countless driver assistance systems. In these systems, earlier detection of a driver's maneuver intentions offers opportunities to improve driving experience and safety. Especially brake maneuvers are of fundamental importance because they are directly related to the avoidance of potential hazards. Current state-of-the-art brake assistance systems rely on the release speed of accelerator pedal as an indicator whether a brake event is planned. However, this simple and practical algorithm, fails to capture the overall movement pattern of accelerator pedal behaviors and cannot utilize rich information from different vehicle sensors. To address this issue, we propose a novel recurrent neural network architecture for the purpose of brake maneuver prediction. The proposed method exploits the advantages of multiple sensors. Unlike conventional practices where all signals are aggregated to a single neural network, we leverage the confidence of each sensor. We evaluate our approach based on a dataset of 44 drivers, comprising around 500 hours of naturalistic driving data. The evaluation results show that the proposed algorithm outperforms baseline method by large margin.

show abstract

“…Having overlapping images available, state-of-the-art methods rely on feature-based localization, e.g. [9,17,29,30,34]. First, they detect local visual features, such as blobs with SIFT [20], which is well suited for ground images [28].…”

Section: Introductionmentioning

confidence: 99%

Deep Metric Learning for Ground Images

Radhakrishnan,

Schmid,

Scholz

et al. 2021

Preprint

View full text Add to dashboard Cite

Ground texture based localization methods are potential prospects for low-cost, highaccuracy self-localization solutions for robots. These methods estimate the pose of a given query image, i.e. the current observation of the ground from a downward-facing camera, in respect to a set of reference images whose poses are known in the application area. In this work, we deal with the initial localization task, in which we have no prior knowledge about the current robot positioning. In this situation, the localization method would have to consider all available reference images. However, in order to reduce computational effort and the risk of receiving a wrong result, we would like to consider only those reference images that are actually overlapping with the query image. For this purpose, we propose a deep metric learning approach that retrieves the most similar reference images to the query image. In contrast to existing approaches to image retrieval for ground images, our approach achieves significantly better recall performance and improves the localization performance of a state-of-the-art ground texture based localization method.

show abstract

StreetMap - Mapping and Localization on Ground Planes using a Downward Facing Camera

Cited by 16 publications

References 16 publications

Model-Based Parameter Optimization for Ground Texture Based Localization Methods

Model-Based Parameter Optimization for Ground Texture Based Localization Methods

Brake Maneuver Prediction – An Inference Leveraging RNN Focus on Sensor Confidence

Deep Metric Learning for Ground Images

Contact Info

Product

Resources

About