• Smart Media Journal
• /
• v.9 no.4
• /
• pp.9-16
• /
• 2020

Recently, various studies are being conducted to integrate Image Segmentation into smart factory industries and autonomous driving fields. In particular, Image Segmentation systems using deep learning algorithms have been researched and developed enough to learn from large volumes of data with higher accuracy. In order to use image segmentation in the autonomous driving sector, sufficient amount of learning is needed with large amounts of data and the streaming environment that processes drivers' data in real time is important for the accuracy of safe operation through highways and child protection zones. Therefore, we proposed a novel DFCN algorithm that enhanced existing FCN algorithms that could be applied to various road environments, demonstrated that the performance of the DFCN algorithm improved 1.3% in terms of "loss" value compared to the previous FCN algorithms. Moreover, the proposed DFCN algorithm was applied to the existing U-Net algorithm to maintain the information of frequencies in the image to produce better results, resulting in a better performance than the classical FCN algorithm in the autonomous environment.

• Proceedings of the KSRS Conference
• /
• 2006.03a
• /
• pp.173-177
• /
• 2006

고해상도의 위성영상이 수집됨에 따라 이에서 지형지물을 자동 추출하려는 분야가 점차 중요해지고 있다. 지형지물을 수작업을 거치지 않고 추출하는 방법의 연구 중에서는 지형지물의 경계추출을 기반으로 하는 방법이 많이 이용된다. 그러나 일반적으로 추출된 지형지물 경계의 선형이 왜곡된 형태를 갖으며, 지형지물의 실제 경계의 형태와는 다소 차이가 있는 결과를 보인다. 이 연구에서는 선형이 굴곡이 있는 지형지물 중 건축물의 경계를 인접한 도로 정보를 이용하여 실제의 경계 형태에 가깝게 보정하는 방법을 검토하였다. context 정보로서 이것을 이용한다면 블록 내의 건축물의 경계를 보다 규칙적으로 정비할 수 있을 것이다. 이 연구에서는 이러한 전제를 가지고 도로로 분할된 블록 내 건축물 경계를 도로의 방향을 기반으로 보정하는 방법을 제안하였다.

• The Journal of Society for e-Business Studies
• /
• v.26 no.1
• /
• pp.93-106
• /
• 2021

Many recent studies on deep learning-based road crack detection have shown significantly more improved performances than previous works using algorithm-based conventional approaches. However, many deep learning-based studies are still focused on classifying the types of cracks. The classification of crack types is highly anticipated in that it can improve the crack detection process, which is currently relying on manual intervention. However, it is essential to calculate the severity of the cracks as well as identifying the type of cracks in actual pavement maintenance planning, but studies related to road crack detection have not progressed enough to automated calculation of the severity of cracks. In order to calculate the severity of the crack, the type of crack and the area of the crack in the image must be identified together. This study deals with a method of using Mobilenet-SSD that is deep learning-based object detection techniques to effectively automate the simultaneous detection of crack types and crack areas. To improve the accuracy of object-detection for road cracks, several experiments were conducted to combine the U-Net for automatic segmentation of input image and object-detection model, and the results were summarized. As a result, image masking with U-Net is able to maximize object-detection performance with 0.9315 mAP value. While referring the results of this study, it is expected that the automation of the crack detection functionality on pave management system can be further enhanced.

• Journal of KIISE
• /
• v.43 no.1
• /
• pp.40-46
• /
• 2016

Digital road map generation is primarily based on artificial satellite photographing or in-site manual survey work. Therefore, these map generation procedures require a lot of time and a large budget to create and update road maps. Consequently, people have tried to develop automated map generation systems using GPS trajectory data sets obtained by public vehicles. A fundamental problem in this road generation procedure involves the extraction of representative trajectory such as main roads. Extracting a representative trajectory requires the base data set of piecewise line segments(GPS-trajectories), which have close starting and ending points. So, geometrically similar trajectories are selected for clustering before extracting one representative trajectory from among them. This paper proposes a new divide- and-conquer approach by partitioning the whole map region into regular grid sub-spaces. We then try to find similar trajectories by sweeping. Also, we applied the $Fr{\acute{e}}chet$ distance measure to compute the similarity between a pair of trajectories. We conducted experiments using a set of real GPS data with more than 500 vehicle trajectories obtained from Gangnam-gu, Seoul. The experiment shows that our grid partitioning approach is fast and stable and can be used in real applications for vehicle trajectory clustering.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.18 no.4
• /
• pp.71-79
• /
• 2019

By using convolutional neural networks (CNNs) based on semantic segmentation, road surface damage detection has being studied. In order to generate the CNN model, it is essential to collect the input and the corresponding labeled images. Unfortunately, such collecting pairs of the dataset requires a great deal of time and costs. In this paper, we proposed a road surface damage detection technique based on semi-supervised learning using pseudo labels to mitigate such problem. The model is updated by properly mixing labeled and unlabeled datasets, and compares the performance against existing model using only labeled dataset. As a subjective result, it was confirmed that the recall was slightly degraded, but the precision was considerably improved. In addition, the $F_1-score$ was also evaluated as a high value.

• Journal of Korean Society of Transportation
• /
• v.27 no.4
• /
• pp.207-215
• /
• 2009

Generally, traffic accidents can be influenced by variables driving conditions including geometric, roadside design, and traffic conditions. Under the circumstance, homogeneous roadway segments were firstly identified using typical geometric variables obtained from field data collections in this study. These field data collections were conducted at highways located in several areas having various regional conditions for examples, outside metropolitan city; level and rolling rural areas. Due to many zero cells in crash database, a Zero Inflated Poisson model was used to develop crash prediction model to overestimated results in this study. It was found that EXPO, radius, grade, guardrail, mountainous terrain, crosswalk and bus-stop have statistically significant influence on vehicle to vehicle crashes at rural multi-lane roadway segments.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.1
• /
• pp.129-136
• /
• 2018

In this paper, we propose a new real-time lane detection method that is efficient for road environment. Existing methods have a problem of low reliability under environmental changes. In order to overcome this problem, we emphasize the lane candidate area by using gray level division. And Extracts a straight line component near the lane by using the Hough transform, and generates an ROI for each straight line based on the extracted coordinates. And integrates the generated ROI images. Then, the lane is determined by dividing the object using the dual queue in the ROI image. The proposed method is able to detect lanes even in the environmental change unlike the conventional method. And It is possible to obtain an advantage that the area corresponding to the background such as sky, mountain, etc. is efficiently removed and high reliability is obtained.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.26 no.3D
• /
• pp.367-374
• /
• 2006

This research revealed the operation problems of the current crossing control systems through inspecting and testing the obstacle detection systems and gate control systems for the crossings. To resolve the problems of the crossing control systems, this research developed new algorithms of four-quadrant gate system and obstacle detection systems combing the functions of rasar sensors and magnetic sensors and tested the reliability of the systems. Currently, the obstacle detection systems and gate control systems controls approaching and departing traffic by simply detecting vehicles and obstacles but do not consider traffic movements at the crossings. In addition, they do not make signal cooperation for gate controls. As a result, such inefficient crossing controls result in unsafe gate controls for drivers. Therefore, the newly developed crossing control systems through this study will provide more effective crossing control services with more strengthen information cooperation within control systems. Besides they will help to reduce train crashes at the crossings by gate control systems considering various driving behaviors.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10c
• /
• pp.488-490
• /
• 1998

본 논문에서는 차량 영역 정보를 이용한 교통 혼잡도 측정 시스템을 설계하고 구현한다. 제시한 교통 혼잡도 측정 시스템은 첫째 영역 분할, 둘째 작은 영역의 직사각형화, 셋째 영역의 병합 및 삭제의 세 단계로 나눌 수 있다. 영역 분할 단계에서 획득한 도로 영상을 주어진 임계치에 의해 영역으로 분할한다. 영역 분할후의 영역 정보 중 차량 영역을 추출하는데 영향을 미치지 않는 작은 영역들을 제거하고 영역을 직사각형화하는 단계를 거친다. 이 단계에서 필요없는 많은 작은 영역 정보들을 제거한다. 마지막으로 차선 별로 영역을 병합, 삭제함으로써 각 차선마다 차량 영역 정보를 추출할 수 있다. 본 논문에서는 이러한 차량 영역 정보를 추출하는 방법을 제시하며, 또한 이를 이용한 효과적인 교통 혼잡도 측정 시스템을 소개하고 평가한다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.44 no.5
• /
• pp.28-34
• /
• 2007

A histogram-based road border line extractor is proposed for an efficient road extraction from the high-resolution satellite imagery. The road border lines are extracted from an edge strength map based on the directional histogram difference between the road and the non-road region. The straight and the curved roads are extracted hierarchically from the edge strength map of the original image and the segmented road cluster images, and the road network is constructed based on the connectivity. Unlike the conventional approaches based on the spectral similarity, the proposed road extraction method is more robust to noise because it extracts roads based on the histogram, and is able to extract both the location and the width of roads. In addition, the proposed method can extract roads with various spectral characteristics by identifying the road clusters automatically. Experimental results on IKONOS multi-spectral satellite imagery with high spatial resolution show that the proposed method can extract the straight and the curved roads as well as the accurate road border lines.