Hybrid strategy for traffic light detection by combining classical and self‐learning detectors

Gao, Feng; Wang, Caimei

doi:10.1049/iet-its.2019.0782

Cited by 19 publications

(13 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On that basis, DeepTLR 31 algorithm for traffic lights recognition was proposed, which achieved significantly better performance than manual features without relying on any prior knowledge. With the help of the concise and efficient AlexNet model, Gao and Wang 32 combined deep learning and traditional recognition technology, implemented image segmentation in HSV space, extracted HOG manual features, and merged SVM-based traditional detectors with deep learning detectors. After the iterative development of RCNN to Faster-RCNN, the two-step deep learning algorithm has made significant progress in image recognition accuracy and computational efficiency.…”

Section: Deep Learning-based Traffic Lights Recognitionmentioning

confidence: 99%

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

Zeng

Liu³

et al. 2021

International Journal of Distributed Sensor Networks

View full text Add to dashboard Cite

Image recognition is susceptible to interference from the external environment. It is challenging to accurately and reliably recognize traffic lights in all-time and all-weather conditions. This article proposed an improved vision-based traffic lights recognition algorithm for autonomous driving, integrating deep learning and multi-sensor data fusion assist (MSDA). We introduce a method to obtain the best size of the region of interest (ROI) dynamically, including four aspects. First, based on multi-sensor data (RTK BDS/GPS, IMU, camera, and LiDAR) acquired in a normal environment, we generated a prior map that contained sufficient traffic lights information. And then, by analyzing the relationship between the error of the sensors and the optimal size of ROI, the adaptively dynamic adjustment (ADA) model was built. Furthermore, according to the multi-sensor data fusion positioning and ADA model, the optimal ROI can be obtained to predict the location of traffic lights. Finally, YOLOv4 is employed to extract and identify the image features. We evaluated our algorithm using a public data set and actual city road test at night. The experimental results demonstrate that the proposed algorithm has a relatively high accuracy rate in complex scenarios and can promote the engineering application of autonomous driving technology.

show abstract

Section: Deep Learning-based Traffic Lights Recognitionmentioning

confidence: 99%

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

Zeng

Liu³

et al. 2021

International Journal of Distributed Sensor Networks

View full text Add to dashboard Cite

show abstract

“…It includes three parts: Part 1) Extraction of candidate regions; Part 2) Two object detectors denoted by "Classical Detector" and "Learning Detector"; Part 3) Scheduling logic to combine the two detectors in Part 2. There are many useful candidate region extraction methods, so we can choose different methods according to different tasks directly, for example, using color information to extract candidate regions of traffic lights [37], using sliding widow method to extract candidate regions of pedestrian [38], using region proposal networks to extract multi-class candidate regions [33], and so on.…”

Section: Combined Object Detection Systemmentioning

confidence: 99%

“…Compared with these disadvantages, the known features with explicit physical meaning are used by the classical method to detect object. In such aspects as scalability, expansibility and interpretability, the classical method is a beneficial supplement to deep learning [36]. So a combined detection method by using both classical and deep learning algorithms is presented in this paper to make full use of the advantages of two kinds of detection methods to achieve a better performance.…”

Section: Introductionmentioning

confidence: 99%

A Combined Object Detection Method With Application to Pedestrian Detection

Gao

Wang

2020

IEEE Access

Self Cite

View full text Add to dashboard Cite

Object detection plays an important role in automatic driving systems. Considering the characteristics of classical and deep learning algorithms, a fusion logic is proposed to combine the advantages of these two kinds of object detectors. The relationship of detection performance among different detectors is established theoretically. According to the established theoretical relationship, the improvement of detection performance by fusion is further studied numerically. Furthermore, an optimization method is proposed to guide the design of the sub-detectors to achieve a better comprehensive performance. The effectiveness of this combined approach is validated by application to the detection of pedestrian, in which a support vector machine trained by the HOG feature of pedestrian is adopted as the classical detector and a comparatively simple transfer convolutional neural network (CNN) based on AlexNet structure acts as the deep learning detector. Several comparative tests with the classical and CNN detectors on the training dataset and other totally different dataset have been conducted to show the advantage of the combined one in ensuring detection performance with simpler network and adaptability to new application conditions.

show abstract

“…Such competitions catch the public's attention and try to show the maturity of automatic driving technologies, but only typical use scenarios are considered. It is far away to ensure the reliability and performance of the automatic driving system (ADS), which relates to traffic safety directly and are influenced by uncontrolled traffic environments [6,7].…”

Section: Introductionmentioning

confidence: 99%

Evolution test by improved genetic algorithm with application to performance limit evaluation of automatic parallel parking system

Gao

Zhang

Han

et al. 2021

IET Intelligent Trans Sys

Self Cite

View full text Add to dashboard Cite

Performance limit evaluation of automatic driving system before putting into the market is critical for driving safety. The evolution test by genetic algorithm (GA) is a method by iteratively generating new test scenarios according to the last test results. To avoid its blind search for better efficiency, a scenario complexity index is proposed to measure the test effectiveness indirectly and guide the evolution process under the assumption that a complex scenario is more challenging to realise automatic driving. The traditional crossover and mutation operators are modified to generate more complex scenarios to improve the test efficiency. The advantage of the improved crossover/mutation operators in increasing the offspring's scenario complexity index is analysed in theory. Moreover, the influence of the design parameters on the evolution test process and the global convergence are also discussed. The new evolution test by this improved GA has been applied to find the collision condition of a parallel automatic parking system to validate its effectiveness.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Hybrid strategy for traffic light detection by combining classical and self‐learning detectors

Cited by 19 publications

References 30 publications

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

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

A Combined Object Detection Method With Application to Pedestrian Detection

Evolution test by improved genetic algorithm with application to performance limit evaluation of automatic parallel parking system

Contact Info

Product

Resources

About