• Transactions of the Korean Society of Automotive Engineers
• /
• v.22 no.3
• /
• pp.50-59
• /
• 2014

This paper proposes a method of localization of vehicle especially the horizontal position for the purpose of recognizing the driving lane. Through tracking road signs, the relative position between the vehicle and the sign is calculated and the absolute position is obtained using the known information from the regulation for installation. The proposed method uses Kalman filter for road sign tracking and analyzes the motion using the pinhole camera model. In order to classify the road sign, ORB(Oriented fast and Rotated BRIEF) features from the input image and DB are matched. From the absolute position of the vehicle, the driving lane is recognized. The Experiments are performed on videos from the highway driving and the results shows that the proposed method is able to compensate the common GPS localization errors.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2004.11a
• /
• pp.373-380
• /
• 2004

GIS was introduced to spatial information of complex and various facility systematically and scientifically. And, recently, new trend has been spread by Mobile GIS owing to its fast development of information and communication technology However, these Mobile GIS techniques were introduced in maintenances of various facilities. Due to its mismanagement, the use became insufficient in facilities causing database creation to require realism and mobility. Therefore, in this research, to road facility of various facilities, it deduced improving the work process by using a Mobile GIS to analyze work process of the road facility database creation. And, through an experiment, by comparing and analyzing road facilities database creation process by existing method and improving method, could improvement of road facility database creation using a Mobile GIS.

• Journal of the Korea Society of Computer and Information
• /
• v.20 no.7
• /
• pp.25-32
• /
• 2015

In this paper, we propose a license plate detection method of running cars in various road images. The proposed method first classifies the road image into day and night images to improve detection accuracy, and then the tail-light regions are detected by finding red color areas in RGB color space. The candidate regions of the license plate areas are detected by using symmetrical property, size, width and variance of the tail-light regions, and to find the license plate areas of the various sizes the morphological operations with adaptive size structuring elements are applied. Finally, the plate area is verified and confirmed with the geometrical and image features of the license plate areas. The proposed method was tested with the various road images and the detection rates (precisions) of 84.2% of day images and 87.4% of night images were achieved.

• Spatial Information Research
• /
• v.3 no.1
• /
• pp.55-63
• /
• 1995

The database design, which logically sets the base structure and orga¬nization of the database, is performed by considering the users requirement, the relations between various data, and the relations between data and application field.The road network data must be created to have geometrical topological structure, because various data elements are needed to recognize the state of each section and to relate between data element. In this study, we propose a method of creating the road network database for an integrated road management system.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.3
• /
• pp.861-880
• /
• 2023

Semantic segmentation of road scene is the key technology of autonomous driving, and the improvement of convolutional neural network architecture promotes the improvement of model segmentation performance. The existing convolutional neural network has the simplification of learning knowledge and the complexity of the model. To address this issue, we proposed a road scene semantic segmentation algorithm based on multi-task collaborative learning. Firstly, a depthwise separable convolution atrous spatial pyramid pooling is proposed to reduce model complexity. Secondly, a collaborative learning framework is proposed involved with saliency detection, and the joint loss function is defined using homoscedastic uncertainty to meet the new learning model. Experiments are conducted on the road and nature scenes datasets. The proposed method achieves 70.94% and 64.90% mIoU on Cityscapes and PASCAL VOC 2012 datasets, respectively. Qualitatively, Compared to methods with excellent performance, the method proposed in this paper has significant advantages in the segmentation of fine targets and boundaries.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.334-335
• /
• 2021

In this paper, the distance traveled by a vehicle in an area where a part of the road is cut is measured using the motion of a parabola. Here, when a car running at a constant speed passes over the edge of a broken road, how far it goes from the edge of the road to fall was calculated.

• Spatial Information Research
• /
• v.10 no.3
• /
• pp.469-480
• /
• 2002

Nowadays, digital maps of 1:5,000 scale are used to plan and review far road alignment design. However, the updating and modifying period of digital maps is not so harmonious as topographical changes caused by rapid developments can be reflected in digital maps, the different areas between real surface and digital map can be found easily. This research is aimed to suggest that the use of high resolution satellite image is effective way to get latest topographical information for road alignment design about wide region. IKONOS satellite images were geometrically corrected, and the road alignment data previously designed by traditional procedure were overlapped on the satellite images. As a result, the satellite image maps clearly described wrong road alignment, and modification of road alignment could be accomplished adequately By these procedures, road alignment design was Improved in quality, and could be reasonable and economic design to prevent modification that would be happened in the next step of practical plan. For the geometric correction method of IKONOS images, Thin Plate Spline(TPS) transformation with large number of ground control points, as well as ortho rectification, was effective.

• Korean Journal of Agricultural Science
• /
• v.37 no.3
• /
• pp.471-479
• /
• 2010

Forest roads of 16,424km have been constructed as infrastructure for efficient management of forest. The demand of forest road have been also increased steadily with SOC conception for forest management and wood production. But, accuracy verification by completion drawing of forest road needed aspects extration of geographic information to sound like forest road construction and completion drawing. However, verification for completion drawing has not ascertained. This study carried out the evaluation for position accuracy about constructed forest road in Chungcheongnam-do for evaluating horizontal position accuracy of completion drawing of forest road. In result, first of distance of completion drawing and real route designed completion drawing longer than the real route as Gongju 83m, Seosan 66m, Nonsan 27m and Dangjin 19m, respectively. Second, RMSE by point-correspondence was 11m~14.7m, buffering analysis appeared difference of 18~24m. Finally, index of shape was the similar completion and real route through 6.5~7.4 and data information of forest road corresponds to be perfect. For such reasons, the existing completion drawings have a problem that it cannot use graphic information for drawing digital map according to the regulation, and there is an urgent need for improvement to solve this problem in the process of design and construction.

• International Journal of Highway Engineering
• /
• v.17 no.2
• /
• pp.107-122
• /
• 2015

PURPOSES : The use of virtual driving tests to determine actual road driving behavior is increasing. However, the results indicate a gap between real and virtual driving under same road conditions road based on ergonomic factors, such as anxiety and speed. In the future, the use of virtual driving tests is expected to increase. For this reason, the purpose of this study is to analyze the gap between real and virtual driving on same road conditions and to use a calibration formula to allow for higher reliability of virtual driving tests. METHODS : An intelligent driving recorder was used to capture real driving. A driving simulator was used to record virtual driving. Additionally, a virtual driving map was made with the UC-Win/Road software. We gathered data including geometric structure information, driving information, driver information, and road operation information for real driving and virtual driving on the same road conditions. In this study we investigated a range of gaps, driving speeds, and lateral positions, and introduced a calibration formula to the virtual record to achieve the same record as the real driving situation by applying the effects of the main causes of discrepancy between the two (driving speed and lateral position) using a linear regression model. RESULTS: In the virtual driving test, driving speed and lateral position were determined to be higher and bigger than in the real Driving test, respectively. Additionally, the virtual driving test reduces the concentration, anxiety, and reality when compared to the real driving test. The formula includes four variables to produce the calibration: tangent driving speed, curve driving speed, tangent lateral position, and curve lateral position. However, the tangent lateral position was excluded because it was not statistically significant. CONCLUSIONS: The results of analyzing the formula from MPB (mean prediction bias), MAD (mean absolute deviation) is after applying the formula to the virtual driving test, similar to the real driving test so that the formula works. Because this study was conducted on a national, two-way road, the road speed limit was 80 km/h, and the lane width was 3.0-3.5 m. It works in the same condition road restrictively.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.6
• /
• pp.121-132
• /
• 2018

Weather is an important factor affecting roadway transportation in many aspects such as traffic flow, driver 's driving patterns, and crashes. This study focuses on the relationship between weather and road surface condition and develops a model to estimate the road surface condition using machine learning. A road surface sensor was attached to the probe vehicle to collect road surface condition classified into three categories as 'dry', 'moist' and 'wet'. Road geometry information (curvature, gradient), traffic information (link speed), weather information (rainfall, humidity, temperature, wind speed) are utilized as variables to estimate the road surface condition. A variety of machine learning algorithms examined for predicting the road surface condition, and a two - stage classification model based on 'Random forest' which has the highest accuracy was constructed. 14 days of data were used to train the model and 2 days of data were used to test the accuracy of the model. As a result, a road surface state prediction model with 81.74% accuracy was constructed. The result of this study shows the possibility of estimating the road surface condition using the existing weather and traffic information without installing new equipment or sensors.