Robust Road Detection and Tracking in Challenging Scenarios Based on Markov Random Fields With Unsupervised Learning

Guo, Chunzhao; Mita, Seiichi; McAllester, David

doi:10.1109/tits.2012.2187896

Cited by 64 publications

(34 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…1a) usually perform pixel-level evaluation in the perspective space. Metrics include the classical true positive (TP) and false positive (FP) rates on the pixel/patch level [20], [21], [22], the accuracy [6] as well as precision/recall and the derived F-measure [7], [10], [18]. In order to capture also traffic participants, Alvarez et al [18] propose to incorporate vehicle detections into the evaluation measure.…”

Section: Related Workmentioning

confidence: 99%

A new performance measure and evaluation benchmark for road detection algorithms

Fritsch

Kühnl

Geiger

2013

16th International IEEE Conference on Intelligent Transportation Systems (ITSC 2013)

566

317

View full text Add to dashboard Cite

Abstract-Detecting the road area and ego-lane ahead of a vehicle is central to modern driver assistance systems. While lane-detection on well-marked roads is already available in modern vehicles, finding the boundaries of unmarked or weakly marked roads and lanes as they appear in inner-city and rural environments remains an unsolved problem due to the high variability in scene layout and illumination conditions, amongst others. While recent years have witnessed great interest in this subject, to date no commonly agreed upon benchmark exists, rendering a fair comparison amongst methods difficult.In this paper, we introduce a novel open-access dataset and benchmark for road area and ego-lane detection. Our dataset comprises 600 annotated training and test images of high variability from the KITTI autonomous driving project, capturing a broad spectrum of urban road scenes. For evaluation, we propose to use the 2D Bird's Eye View (BEV) space as vehicle control usually happens in this 2D world, requiring detection results to be represented in this very same space. Furthermore, we propose a novel, behavior-based metric which judges the utility of the extracted ego-lane area for driver assistance applications by fitting a driving corridor to the road detection results in the BEV. We believe this to be important for a meaningful evaluation as pixel-level performance is of limited value for vehicle control. State-of-the-art road detection algorithms are used to demonstrate results using classical pixellevel metrics in perspective and BEV space as well as the novel behavior-based performance measure. All data and annotations are made publicly available on the KITTI online evaluation website in order to serve as a common benchmark for road terrain detection algorithms.

show abstract

Section: Related Workmentioning

confidence: 99%

A new performance measure and evaluation benchmark for road detection algorithms

Fritsch

Kühnl

Geiger

2013

16th International IEEE Conference on Intelligent Transportation Systems (ITSC 2013)

566

317

View full text Add to dashboard Cite

show abstract

“…The third type of road detection works use ego-vehicles with onboard cameras with driver assistance systems or UGVs autonomous navigation systems. A substantial amount of work [7], [25]- [29] have been done in this area. Since the focus of this paper is on road detection and tracking using low-/mid- altitude UAVs, we only give a review of the most related works in this area.…”

Section: IImentioning

confidence: 99%

Tracking and Map Formatting of Road Using Unmanned Ground Vehicle

Kumar¹

2017

IJRASET

View full text Add to dashboard Cite

An unmanned ground vehicle (UGV) has many applications in a variety of fields. Detection and tracking of a specific road in UGV videos play an important role in automatic UGV navigation, traffic monitoring, and is very helpful for constructing road networks for modeling and simulation. In this paper, an efficient road detection and tracking framework in UGV videos is proposed. In particular, a graph-cut-based detection approach is given to accurately extract a specified road region during the initialization stage and in the middle of tracking process, and a fast homography-based tracking scheme is developed to automatically tracking road areas. The high efficiency of our framework is attributed to two aspects: the road detection is performed only when needed and most work in locating the road is rapidly done via very fast homography-based tracking. They use UGVs in autonomous navigation to follow roads/rivers, oil-gas pipeline inspection, and traffic parameters measurements. UAVs equipped with cameras are viewed as a low-cost platform that can provide efficient data acquisition methods for intelligent transport systems. With the increasing vehicles usage and their traffic management demands. this kind of platform becomes attractively popular. Conventional traffic data collection [5] relying on fixed infrastructure is only limited to a local region and, thus, it is expensive and labor intensive to monitor traffic activities across broad areas. In comparison, UAV has advantages, including: There is a low cost to monitor over long distances; it is flexible for flying across broad spatial and temporal scales; and it is capable of carrying various types of sensors to collect abundant data. Detection in deep about road areas can provide users the regions of interest for further navigation, detection and data collection procedures, benefiting their efficiencies and accuracies. In the previous works of road detection and tracking, most approaches use the color (texture) and/or structure (geometry) properties of roads. Among them, the combination of road color and boundary information have achieved more robust and ac-curate results than using only one of them in road detection, as shown in the work [6], [7]. Therefore, we are paying a note of using both types of information. Because real time is required in many UAV-based applications, our major target is how to effectively combine both types of information for road detection/ tracking in an efficient way. There are two rules that make one integrated framework efficient. First, each and every component of the framework should be fast. Second, if one component is faster than the others in achieving the same purpose, fastest component should be prioritized as much as possible. we follow the above mentioned two rules to make the framework fast. our framework includes two components: road detection, road tracking. In road detection, we utilize the GraphCut algorithm [8] considering its efficiency and powerful segmentation performance in 2-D color images. In road tr...

show abstract

“…In this case, we can only use two-dimensional consecutive frames of a video sequence. The use of adjacent frames is described in articles [14,15]. Tracking traffic sign on adjacent frames can not only increase the confidence in the correct detection, but also reduce the computational complexity of the algorithm by reducing the search area in the adjacent frames.…”

Section: Detection and Trackingmentioning

confidence: 99%

Traffic Signs Detection Using Tracking with Prediction

Yakimov

2016

E-Business and Telecommunications

View full text Add to dashboard Cite

Abstract. This paper proposes an efficient algorithm for real-time traffic sign detection. The article considers the practicability of using HSV color space to extract the red color. An algorithm to remove noise to improve the accuracy and speed of detection was developed. A modified Generalized Hough transform is then used to detect traffic signs. The current velocity of a vehicle is used to predict the sign's location in the adjacent frames in a video sequence. Finally, the detected objects are being classified. The detection and classification of road signs algorithms are implemented using CUDA and operate in real time on an Android device. The developed algorithms have been tested using real scene images and the German Traffic Sign Detection Benchmark (GTSDB) dataset and showed efficient results.

show abstract

Robust Road Detection and Tracking in Challenging Scenarios Based on Markov Random Fields With Unsupervised Learning

Cited by 64 publications

References 41 publications

A new performance measure and evaluation benchmark for road detection algorithms

A new performance measure and evaluation benchmark for road detection algorithms

Tracking and Map Formatting of Road Using Unmanned Ground Vehicle

Traffic Signs Detection Using Tracking with Prediction

Contact Info

Product

Resources

About