Multi-Stage CNN-Based Monocular 3D Vehicle Localization and Orientation Estimation

Babolhavaeji, Ali; Shykholeslami, Mohammad Ali Fanaei

doi:10.1109/csci51800.2020.00295

Cited by 3 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Like satellite images, highresolution aerial images have also been used for vehicle localization using CNN [105]. Another study on CNN-based vehicle localization is presented in [106], which uses a bird'seye view elevation map and the deep representation of object features. [107] employed the CNN for camera-Radar sensor fusion as well as for vehicle corner detection.…”

Section: A Neural Networkmentioning

confidence: 99%

A Survey on Localization for Autonomous Vehicles

Kumar,

Muhammad

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

Section: A Neural Networkmentioning

confidence: 99%

A Survey on Localization for Autonomous Vehicles

Kumar,

Muhammad

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The main concept is divided into the following parts: (1) use a 2D object detector and (2) extract features from the cropped vehicles than using an attention mechanism for estimation of the 3D predictions. Authors of the paper [ 11 ] proposed to use the bird eye view to estimate the depth values and use the results to estimate the 3D detection from 2D detection, assuming that rotation will be only around one axis (yaw angle), making it weak against terrain scenes (mountains, hills, or even bridges).…”

Section: Related Workmentioning

confidence: 99%

3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks

Kashevnik

Ali

2022

Sensors

View full text Add to dashboard Cite

In this paper, we present a two stages solution to 3D vehicle detection and segmentation. The first stage depends on the combination of EfficientNetB3 architecture with multiparallel residual blocks (inspired by CenterNet architecture) for 3D localization and poses estimation for vehicles on the scene. The second stage takes the output of the first stage as input (cropped car images) to train EfficientNet B3 for the image recognition task. Using predefined 3D Models, we substitute each vehicle on the scene with its match using the rotation matrix and translation vector from the first stage to get the 3D detection bounding boxes and segmentation masks. We trained our models on an open-source dataset (ApolloCar3D). Our method outperforms all published solutions in terms of 6 degrees of freedom error (6 DoF err).

show abstract

Next-gen Urban Buses: Autonomy and Connectivity

Hsu

2024

View full text Add to dashboard Cite

<div class="section abstract"><div class="htmlview paragraph">The deployment of autonomous urban buses brings with it the hope of addressing concerns associated with safety and aging drivers. However, issues related autonomous vehicle (AV) positioning and interactions with road users pose challenges to realizing these benefits. This report covers unsettled issues and potential solutions related to the operation of autonomous urban buses, including the crucial need for all-weather localization capabilities to ensure reliable navigation in diverse environmental conditions. Additionally, minimizing the gap between AVs and platforms during designated parking requires precise localization.</div><div class="htmlview paragraph"><b>Next-gen Urban Buses: Autonomy and Connectivity</b> addresses the challenge of predicting the intentions of pedestrians, vehicles, and obstacles for appropriate responses, the detection of traffic police gestures to ensure compliance with traffic signals, and the optimization of traffic performance through urban platooning—including the need for advanced communication and coordination technology to maintain stability and reliability in high-traffic scenarios.</div><div class="htmlview paragraph"><a href="https://www.sae.org/publications/edge-research-reports" target="_blank">Click here to access the full SAE EDGE</a><sup>TM</sup><a href="https://www.sae.org/publications/edge-research-reports" target="_blank"> Research Report portfolio.</a></div></div>

show abstract

Multi-Stage CNN-Based Monocular 3D Vehicle Localization and Orientation Estimation

Cited by 3 publications

References 27 publications

A Survey on Localization for Autonomous Vehicles

A Survey on Localization for Autonomous Vehicles

3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks

Next-gen Urban Buses: Autonomy and Connectivity

Contact Info

Product

Resources

About