• Korean Journal of Artificial Intelligence
• /
• v.11 no.4
• /
• pp.1-8
• /
• 2023

In this paper, we propose a black ice detection platform framework using Convolutional Neural Networks (CNNs). To overcome black ice problem, we introduce a real-time based early warning platform using CNN-based architecture, and furthermore, in order to enhance the accuracy of black ice detection, we apply a multi-scale dilation convolution feature fusion (MsDC-FF) technique. Then, we establish a specialized experimental platform by using a comprehensive dataset of thermal road black ice images for a training and evaluation purpose. Experimental results of a real-time black ice detection platform show the better performance of our proposed network model compared to conventional image segmentation models. Our proposed platform have achieved real-time segmentation of road black ice areas by deploying a road black ice area segmentation network on the edge device Jetson Nano devices. This approach in parallel using multi-scale dilated convolutions with different dilation rates had faster segmentation speeds due to its smaller model parameters. The proposed MsCD-FF Net(2) model had the fastest segmentation speed at 5.53 frame per second (FPS). Thereby encouraging safe driving for motorists and providing decision support for road surface management in the road traffic monitoring department.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.3
• /
• pp.17-22
• /
• 2023

In this paper, we propose a black ice detection system using Convolutional Neural Networks (CNNs). Black ice poses a serious threat to road safety, particularly during winter conditions. To overcome this problem, we introduce a CNN-based architecture for real-time black ice detection with an encoder-decoder network, specifically designed for real-time black ice detection using thermal images. To train the network, we establish a specialized experimental platform to capture thermal images of various black ice formations on diverse road surfaces, including cement and asphalt. This enables us to curate a comprehensive dataset of thermal road black ice images for a training and evaluation purpose. Additionally, in order to enhance the accuracy of black ice detection, we propose a multi-scale dilation convolution feature fusion (MsDC-FF) technique. This proposed technique dynamically adjusts the dilation ratios based on the input image's resolution, improving the network's ability to capture fine-grained details. Experimental results demonstrate the superior performance of our proposed network model compared to conventional image segmentation models. Our model achieved an mIoU of 95.93%, while LinkNet achieved an mIoU of 95.39%. Therefore, it is concluded that the proposed model in this paper could offer a promising solution for real-time black ice detection, thereby enhancing road safety during winter conditions.

• Spatial Information Research
• /
• v.17 no.2
• /
• pp.213-222
• /
• 2009

Finding exclusive closest pairs in road network is very useful to real applications such as, for example, finding a closest pair between a passenger and a nearby taxi in a road network. Few studies, however, have been interested in this problem. To match two close moving objects exclusively, one object must belong to only one result pair. Because moving objects in a road network change their position continuously, it is necessary to monitor closest pair results. In this paper, we propose a methodology to monitor k exclusive closest pairs via a road network. Proposed method only updates the results which are influenced by objects' movement. We evaluated the performance of the proposed method with various real road network data. The results show that our method produces better accuracy than normal batch processing methods.

• Journal of Electrical Engineering and Technology
• /
• v.10 no.4
• /
• pp.1836-1842
• /
• 2015

On road networks, the clustering of moving objects is important for traffic monitoring and routes recommendation. The existing schemes find out density route by considering the number of vehicles in a road segment. Since they don’t consider the features of each road segment such as width, length, and directions in a road network, the results are not correct in some real road networks. To overcome such problems, we propose a clustering method for congested routes discovering from the trajectories of moving objects on road networks. The proposed scheme can be divided into three steps. First, it divides each road network into segments with different width, length, and directions. Second, the congested road segments are detected through analyzing the trajectories of moving objects on the road network. The saturation degree of each road segment and the average moving speed of vehicles in a road segment are computed to detect the congested road segments. Finally, we compute the final congested routes by using a clustering scheme. The experimental results showed that the proposed scheme can efficiently discover the congested routes in different directions of the roads.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.1
• /
• pp.193-202
• /
• 2021

The update of a high-precision map was carried out by modifying the geometric information using ortho-images or point-cloud data as the source data and then reconstructing the relationship between the spatial objects. These series of processes take considerable time to process the geometric information, making it difficult to apply real-time route planning to a vehicle quickly. Therefore, this study proposed a method to update the road network for route planning using a graph data structure and storage type of graph data structure considering the characteristics of the road network. The proposed method was also reviewed to assess the feasibility of real-time route information transmission by applying it to actual road data.

• Management Science and Financial Engineering
• /
• v.12 no.2
• /
• pp.87-110
• /
• 2006

In a hierarchical road network, all roads can be classified according to their attributes such as speed limit, number of lanes, etc. By splitting the whole road network into the subnetworks of the highlevel and low-level roads, we can reduce the size of the network to be calculated at once, and find a path in the way that drivers usually adopt when searching out a travel route. To exploit the hierarchical property of road networks, we define a convenient path and propose an algorithm for finding convenient paths. We introduce a parameter indicating the driver's tolerance to the difference between the length of a convenient path and that of a shortest convenient path. From this parameter, we can determine how far we have to search for the entering and exiting gateway. We also propose some techniques for reducing the number of pairs of entries and exits to be searched in a road network. A result of the computational experiment on a real road network is given to show the efficiency of the proposed algorithm.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2009.05a
• /
• pp.199-202
• /
• 2009

The optimal path on real road network has been changed by traffic flow of roads frequently. Therefore a path finding system to find the optimal path on real network should consider traffic flow of roads that is changed on real time. The most of existing path finding methods do not consider traffic flow of roads and do not also perform efficiently if they use traffic information. In this paper, we propose an abstraction method of real road network based on the Terminal Based Navigation System (TBNS) with technique such as TPEG. TBNS can be able to provides quality of path better than before as using traffic information that is transferred by TPEG. The proposed method is to abstract real network as simple graph in order to use traffic information. It is composed boundary nodes based on real nodes, all boundary nodes that have the same of connection are merged together. The result of path finding on an abstract graph diminishes the search space.

• Journal of the Korea Computer Graphics Society
• /
• v.17 no.4
• /
• pp.23-30
• /
• 2011

We propose an intelligent ribbon road network for automatic vehicle simulation, and a real-time algorithm for large-scale, realistic traffic simulation based on artificial energy functions. Our method reconstructs a road network automatically from both GIS (Geographic Information System) real-world data and synthetic models. Such automatic road network helps us to easily simulate almost every possible scenario such as intersections, ramps, etc. In order to simulate agents' movement, we design car-environment interaction energy and car-car interaction energy functions. Car agents move along the road network according to the proposed energy functions while avoiding collisions with other car agents.

• Journal of Korea Spatial Information System Society
• /
• v.10 no.1
• /
• pp.29-40
• /
• 2008

With the spread of mobile devices and advances in communication techknowledges, the needs of application which uses the movement patterns of moving objects in history trajectory data of moving objects gets Increasing. Especially, to design public transportation route or road network of the new city, we can use the similar patterns in the trajectories of moving objects that move on the spatial network such as road and railway. In this paper, we propose a spatio-temporal similar trajectory search algorithm for moving objects on road network. For this, we define a spatio-temporal similarity measure based on the real road network distance and propose a grid-based index structure for similar trajectory search. Finally, we analyze the performance of the proposed similar trajectory search algorithm in order to show its efficiency.

• Journal of IKEEE
• /
• v.25 no.4
• /
• pp.756-761
• /
• 2021

Dynamic route guidance (DRG) finds the fastest path from a source to a destination location considering the real-time congestion information. In Korea, the traffic state information is available by the public transportation data (PTD) which is indexed on top of the node-link map (NLM). While the NLM is the authoritative low-detailed road network for major roads only, the OpenStreetMap road network (ORN) supports not only a high-detailed road network but also a few open-source routing engines, such as OSRM and Valhalla. In this paper, we propose a DRG framework based on road network matching between the NLM and ORN. This framework regularly retrieves the NLM-indexed PTD to construct a historical speed profile which is then mapped to ORN. Next, we extend the Valhalla routing engine to support dynamic routing based on the historical speed profile. The numerical results at the Yeoui-do island with collected 11-month PTD show that our DRG framework reduces the travel time up to 15.24 % and improves the estimation accuracy of travel time more than 5 times.