• Journal of Korea Society of Industrial Information Systems
• /
• v.18 no.2
• /
• pp.19-25
• /
• 2013

In this paper, we propose an approach to tracking road regions from video sequences. The proposed method segments and tracks road regions by utilizing the prior information from the result of the previous frame. For the efficiency of the system, we have a simple assumption that the road region is usually shown in the lower part of input images so that lower 60% of input images is set to the region of interest(ROI). After initial segmentation using flood-fill algorithm, we merge neighboring regions based on color similarity measure. The previous segmentation result, in which seed points for the successive frame are extracted, is used as prior information to segment the current frame. The similarity between the road region of the previous frame and that of the current frame is measured by the modified Jaccard coefficient. According to the similarity we refine and track the detected road regions. The experimental results reveal that the proposed method is effective to segment and track road regions in noisy and non-noisy environments.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.6
• /
• pp.87-97
• /
• 2018

Rainfall depth is an important meteorological information. Generally, high spatial resolution rainfall data such as road-level rainfall data are more beneficial. However, it is expensive to set up sufficient Automatic Weather Systems to get the road-level rainfall data. In this paper, we proposed to use deep learning to recognize rainfall depth from road surveillance videos. To achieve this goal, we collected two new video datasets, and proposed a new deep learning architecture named Temporal and Spatial Segment Networks (TSSN) for rainfall depth recognition. Under TSSN, the experimental results show that the combination of the video frame and the differential frame is a superior solution for the rainfall depth recognition. Also, the proposed TSSN architecture outperforms other architectures implemented in this paper.

• Journal of Korean Society for Atmospheric Environment
• /
• v.34 no.5
• /
• pp.735-747
• /
• 2018

Rainfall depth is an important meteorological information. Generally, high spatial resolution rainfall data such as road-level rainfall data are more beneficial. However, it is expensive to set up sufficient Automatic Weather Systems to get the road-level rainfall data. In this paper, we propose to use deep learning to recognize rainfall depth from road surveillance videos. To achieve this goal, we collect a new video dataset and propose a procedure to calculate refined rainfall depth from the original meteorological data. We also propose to utilize the differential frame as well as the optical flow image for better recognition of rainfall depth. Under the Temporal Segment Networks framework, the experimental results show that the combination of the video frame and the differential frame is a superior solution for the rainfall depth recognition. The final model is able to achieve high performance in the single-location low sensitivity classification task and reasonable accuracy in the higher sensitivity classification task for both the single-location and the multi-location case.

• Korean Journal of Remote Sensing
• /
• v.33 no.4
• /
• pp.423-436
• /
• 2017

This paper presents a deep learning-based road segmentation framework from very high-resolution orthophotos. The proposed method uses Deep Convolutional Autoencoders for end-to-end mapping of orthophotos to road segmentations. In addition, a set of post-processing steps were applied to make the model outputs GIS-ready data that could be useful for various applications. The optimization of the model's parameters is explained which was conducted via grid search method. The model was trained and implemented in Keras, a high-level deep learning framework run on top of Tensorflow. The results show that the proposed model with the best-obtained hyperparameters could segment road objects from orthophotos at an average accuracy of 88.5%. The results of optimization revealed that the best optimization algorithm and activation function for the studied task are Stochastic Gradient Descent (SGD) and Exponential Linear Unit (ELU), respectively. In addition, the best numbers of convolutional filters were found to be 8 for the first and second layers and 128 for the third and fourth layers of the proposed network architecture. Moreover, the analysis on the time complexity of the model showed that the model could be trained in 4 hours and 50 minutes on 1024 high-resolution images of size $106{\times}106pixels$, and segment road objects from similar size and resolution images in around 14 minutes. The results show that the deep learning models such as Convolutional Autoencoders could be a best alternative to traditional machine learning models for road segmentation from aerial photographs.

• International Journal of Highway Engineering
• /
• v.16 no.5
• /
• pp.133-142
• /
• 2014

PURPOSES : The objective was to develop the advanced method which could not explain each observation's specific characteristic in the present negative binomial method that results in under-estimation of the standard error(t-value inflation) and affects the confidence of whole derived results. METHODS : This study dealt with traffic accidents occurring within interchange segment on highway main line with RPNB(Random Parameter Negative Binomial) method that enables to take account of heterogeneity. RESULTS : As a result, AADT and lighting installation type on the road were revealed to have random parameter and in terms of other geometric variables, all were derived as fixed parameter(same effect on every segment). Also, marginal effects were adapted to analyze the relative effects on traffic accidents. CONCLUSIONS : This study proves that RPNB method which considers each observation's specific characteristics is better fitted to the accident data with geometrics. Thus, it is recommended that RPNB model or other methods which could consider the heterogeneity needs to be adapted in accident analysis.

• Proceedings of the KSRS Conference
• /
• 2002.10a
• /
• pp.34-39
• /
• 2002

Road information is very important for topographic mapping, transportation application, urban planning and other related application fields. Therefore, automatic detection of road networks from spatial imagery, such as aerial photos and satellite imagery can play a central role in road information acquisition. In this paper, we use least squares correlation matching alone for road center tracking and show that it works. We assumed that (bright) road centerlines would be visible in the image. We further assumed that within a same road segment, there would be only small differences in brightness values. This algorithm works by defining a template around a user-given input point, which shall lie on a road centerline, and then by matching the template against the image along the orientation of the road under consideration. Once matching succeeds, new match proceeds by shifting a matched target window further along road orientation at the target window. By repeating the process above, we obtain a series of points, which lie on a road centerline successively. A 1m resolution IKONOS images over Seoul and Daejeon were used for tests. The results showed that this algorithm could extract road centerlines in any orientation and help in fast and exact he ad-up digitization/vectorization of cartographic images.

• Proceedings of the KSR Conference
• /
• 2011.10a
• /
• pp.1802-1809
• /
• 2011

A distribution method of limited budget for rail line facility improvement is to use investment priorities of rail line segments based on their deterioration. In this paper we present an evaluation method of rail line segment deterioration which can be used to distribute limited budget. Rail line facilities include rail track, crossing, road bed, bridge, tunnel. These facilities deterioration and line shape can affect line segment deterioration. Deterioration evaluation method we present is a weighted sum of each component deterioration scores. The component weight can be obtained from experts using analytic hierarchy process.

• Journal of Korean Society of Transportation
• /
• v.23 no.1
• /
• pp.21-32
• /
• 2005

At present automated speed enforcement system in Korea control overspeed vehicle only in the specific spot. Because the drivers generally recognize the previous stated fact, they reduce a speed only in the establishment location of systems and increase rapidly again as soon as it passes the location. we have known that the rate of traffic risk at the tunnel, bridge and curve road segment is higher than other road section. Therefore, it needs speed control in them. In such a case, it is necessary to establish the automated traffic enforcement system based on the travel time speed of an individual vehicle over a pre defined stretch of road. In this study, the application limit of existing spot overspeed enforcement system was studied through an analysis of traffic flow characteristics in the tunnel, bridge and curve section. Also we found out the optimal distance of segment and the most suitable location to an application of the overspeed vehicle by travel time speed through an analysis of the road structure, traffic condition and accident numbers in the road.

• Journal of Korean Society of Transportation
• /
• v.23 no.7 s.85
• /
• pp.77-86
• /
• 2005

Level of service(LOS) is a quantify measure describing operational conditions within a traffic stream, generally, in terms of such service measures as speed, travel time, freedom to measures, traffic interruptions, comfort and convenience. The LOS is leveled by highway facilities according to measure of effectiveness(MOE) and then used to evaluate performance capacity. The current evaluation of a urban road is performed by only a aspect of traffic operation without any concepts of safety. Therefore, this paper presents a method for evaluation of risk order for urban road with new MOE, user cost analysis, considering both smooth traffic operation(congestion) and traffic safety(accident). The user coat is included traffic accident cast by traffic safety and traffic congestion cost by traffic operation. First of all, a number of traffic accident and accident rate by highway geometric is inferred from urban road traffic accident prediction model (Poul Greibe(2001)) Secondly, a user cost is inferred as traffic accident cast and traffic congestion cost is putting together. Thirdly, a method for evaluation of a urban road is inferred by user cost analysis. Fourthly a accident rate by segment predict with traffic accidents and data related to the accidents in $1996{\sim}1998$ on 11 urban road segments, Gwang-Ju, predicted accident rate. Traffic accident cost predict using predicted accident rate, and, traffic congestion cost predict using predicted average traffic speed(KHCM). Fifthly, a risk order are presented by predicted user cost at each segment in urban roads. Finally, it si compared and evaluated that LOS of 11 urban road segments, Gwang-Ju, by only a aspect of traffic operation without any concepts of safety and risk order by a method for evaluation of urban road in this paper.

• International Journal of Highway Engineering
• /
• v.17 no.5
• /
• pp.75-82
• /
• 2015

PURPOSES : In link matching using vehicular trajectory in a car navigation system, it is technically difficult to match the location of the subject vehicle with a link on an electronic map in the vehicle on a grade separation highway segment consisting of an elevated highway and atgrade highway, because of the overlap of geometric lines of the two highways. This study was initiated to propose a link matching algorithm using the characteristics of vehicular movement of the subject vehicle. METHODS : The selected test site is the highway segment between Jeong-reung IC and Gil-eum IC where the Inner Circulation Road and Jeong-reung-ro run together. To evaluate the proposed algorithm, this study collected the raw packet data of vehicles that drove on the test site. In a simulation environment, link matching was performed using an existing algorithm as well as the proposed algorithm. RESULTS: It was clearly found that the characteristics of vehicular movements are different on the two highways. CONCLUSIONS: The proposed algorithm outperformed the existing algorithm.