Traffic Light Recognition Using Deep Learning and Prior Maps for Autonomous Cars

@article{Possatti2019TrafficLR,
  title={Traffic Light Recognition Using Deep Learning and Prior Maps for Autonomous Cars},
  author={Lucas C. Possatti and R{\^a}nik Guidolini and Vinicius B. Cardoso and Rodrigo Berriel and Thiago Meireles Paix{\~a}o and Claudine Santos Badue and Alberto Ferreira de Souza and Thiago Oliveira-Santos},
  journal={2019 International Joint Conference on Neural Networks (IJCNN)},
  year={2019},
  pages={1-8}
}
Autonomous terrestrial vehicles must be capable of perceiving traffic lights and recognizing their current states to share the streets with human drivers. [] Key Method The process is divided in two phases: an offline phase for map construction and traffic lights annotation; and an online phase for traffic light recognition and identification of the relevant ones. The proposed system was evaluated on five test cases (routes) in the city of Vitória, each case being composed of a video sequence and a prior map…

Figures and Tables from this paper

An improved traffic lights recognition algorithm for autonomous driving in complex scenarios

An improved vision-based traffic lights recognition algorithm for autonomous driving, integrating deep learning and multi-sensor data fusion assist (MSDA) is proposed and the optimal ROI can be obtained to predict the location of traffic lights.

Relevant Traffic Light Localization via Deep Regression

This work proposes a deep regression system with an outliers resilient loss to predict the coordinates of a relevant traffic light in the image plane that can be used as a high-level decisionmaker or an assistant to a cheaper classifier to work on a region of interest.

Simultaneous detection and tracking using deep learning and integrated channel feature for ambint traffic light recognition

Deep learning based detection and object tracking is synthesized to determine the position and color of traffic lights and results indicate that the proposed technique can improve the accuracy and speed of recognition.

Simultaneous detection and tracking using deep learning and integrated channel feature for ambint traffic light recognition

Deep learning based detection and object tracking is synthesized to determine the position and color of traffic lights and results indicate that the proposed technique can improve the accuracy and speed of recognition.

Comparison Study of Traffic Signs Recognition Using Deep Learning Architectures

  • Khaldaa AlawajiR. Hedjar
  • Computer Science
    2022 13th International Conference on Information and Communication Systems (ICICS)
  • 2022
A comparison study of object recognition challenges for self-driving automobiles, in particular for traffic signs and traffic signals detection, using two standard datasets and six different deep CNNs architectures to solve the unbalanced datasets problem.

Traffic Signal Classification with Cost-Sensitive Deep Learning Models

Evaluating several cost-sensitive techniques applicable to deep learning models in traffic signal classification found that the resulting models did not improve accuracy in prediction and may need to be considered for future studies.

Visual Global Localization Based on Deep Neural Netwoks for Self-Driving Cars

A visual global localization system based on Deep Neural Networks (DNNs) for self-driving cars, named DeepVgl (Deep Visual Global Localization), which is able to correctly locate the self- Driving car up to 75% of the time for 0.2 m of accuracy.

Traffic Lights Detection and Recognition with New Benchmark Datasets Using Deep Learning and TensorFlow Object Detection API

Faster R-CNN Inception v2 deep learning model was trained and tested on two different datasets that were prepared and published publicly under variable traffic and climatic conditions in Turkey and it has been shown that the datasets prepared can be developed considering the conditions in other countries and successful results will be obtained.

TL-GAN: Improving Traffic Light Recognition via Data Synthesis for Autonomous Driving

Extensive experiments show that the proposed TL-GAN renders remarkable improvement over the baseline without using the generated data, leading to the state-of-the-art performance in comparison with the competing algorithms that are used for general image synthesis and data imbalance tackling.

VATLD: A Visual Analytics System to Assess, Understand and Improve Traffic Light Detection

This work proposes a visual analytics system, VATLD, equipped with a disentangled representation learning and semantic adversarial learning, to assess, understand, and improve the accuracy and robustness of traffic light detectors in autonomous driving applications.

References

SHOWING 1-10 OF 32 REFERENCES

A deep learning approach to traffic lights: Detection, tracking, and classification

A complete system consisting of a traffic light detector, tracker, and classifier based on deep learning, stereo vision, and vehicle odometry which perceives traffic lights in real-time is proposed.

Deep Convolutional Traffic Light Recognition for Automated Driving

This work developed a unified deep convolutional traffic light recognition system on the basis of the Faster R-CNN architecture, which is able to not only detect traffic lights and classify their state, but also distinguish their type (circle, straight, left, and right).

A Hierarchical Deep Architecture and Mini-batch Selection Method for Joint Traffic Sign and Light Detection

This work presents a deep hierarchical architecture in conjunction with a mini-batch proposal selection mechanism that allows a network to detect both traffic lights and signs from training on separate traffic light and sign datasets to solve the overlapping issue where instances from one dataset are not labelled in the other dataset.

Traffic light recognition exploiting map and localization at every stage

Traffic light recognition in varying illumination using deep learning and saliency map

A convolutional neural network is used to extract and detect features from visual camera images to improve the recognition accuracy of the proposed traffic light recognition algorithm in varied illumination conditions.

Traffic light mapping and detection

Methods for automatically mapping the three dimensional positions of traffic lights and robustly detecting traffic light state onboard cars with cameras are presented.

Traffic Light Detection: A Learning Algorithm and Evaluations on Challenging Dataset

An extensive public database is collected based on footage from US roads, which shows that the learning based detector achieves an AUC of 0.4 and 0.32 for day sequence 1 and 2, respectively, which is more than an order of magnitude better than the two heuristic model-based detectors.

Traffic light mapping, localization, and state detection for autonomous vehicles

This work introduces a convenient technique for mapping traffic light locations from recorded video data using tracking, back-projection, and triangulation, and is the first to account for multiple lights per intersection, which yields superior results by probabilistically combining evidence from all available lights.

Evaluating State-of-the-Art Object Detector on Challenging Traffic Light Data

The state-of-the-art, real-time object detection system You Only Look Once, (YOLO) is applied on the public LISA Traffic Light dataset available through the VIVA-challenge, which contain a high number of annotated traffic lights, captured in varying light and weather conditions.

The DriveU Traffic Light Dataset: Introduction and Comparison with Existing Datasets

The DriveU**driveU is a joint innovation center of the Daimler AG and the University of Ulm Traffic Light Dataset (DTLD), a large-scale dataset consisting of more than 230,000 annotations, which exceeds previous traffic light datasets in size, variance, annotation quality and amount of additional sensor data.