바로가기 및 건너띄기 링크
본문 바로가기
주메뉴 바로가기
KOREA INSTITUTE of CIVIL ENGINEERING and BUILDING TECHNOLOGY home

HOME > Research Information > Research Achievements

Research Achievements

주요연구성과
Titel RRAP, A State-of-the-art Platform for the Automatic Analysis of Road Signs
Manage ICT Convergence and Integration Research Instiute Date 2017.03.03 Hit 2,855

 

"Ensuring the people’s safety and reducing the national budget

by developing automatic road sign recognition and analysis equipment"

 

Project Leader: Chong, Kyu-soo (Research Fellow, ksc@kict.re.kr)

 

 

 
Road signs play the role of directing people to the place where they want to go. Therefore, road signs should be regularly maintained and improved to prevent errors. The vehicle equipped with the Road Sign Recognition and Analysis Vehicle Platform (RRAP) developed in this research collects visual information and automatically recognizes information on the name of a place and the directions indicated on road signs while being driven so as to detect wrong information and to reflect the data onto the road sign database, which enables systematic road sign maintenance. It is expected that the introduction of RRAP will promote the people’s safety and will drastically reduce the national budget.
 

 

 


Faster and more accurate management of the 160k road signs nationwide!

 

There are currently about 160,000 road signs nationwide on highways, national roads, and local roads, which are managed by some 300 road management offices. For the efficient management and improvement of the national road signs, the government has been running the “Road Sign System” since 2001, but it is difficult to renew the newly installed or revised road sign information in real time, and as the acquisition of onsite images and location information for road management and property information entry is performed manually, it is unrealistic to maintain the real-time information.  Every year, road signs are changed due to the construction of new roads, and the government is making every effort to renew the road sign information in real time for the convenience and safety of the people. A re-survey of all the 160,000 road signs nationwide, however, will cost KRW10 billion for the onsite survey and KRW5.2 billion each year for entering and analyzing the information. Furthermore, the acquired data cannot have 100% reliability. Therefore, the introduction of a state-of-the-art equipment that can automatically detect and analyze road signs more accurately and quickly is urgently needed.
 

 


RRAP development process

 

This research aimed to develop an advanced equipment that can quickly and accurately manage and survey the 160,000 road signs nationwide. The research process was largely divided into the development of the road visual information acquisition and classification module, the development of the road sign property information extraction module, the development of the road sign property information analysis module, and the development of the RRAP-road sign system linkage module.
 
In the first phase of the study, the research team developed a technology that collects and processes the visual-spatial information on the road based on the mobile mapping system (MMS), which, when installed in a vehicle, measures the surrounding terrain features while the vehicle is on the road, and extracts property information related to the detection of road signs. Developed in this phase of the research were the multi-sensor calibration1) S/W, the road sign survey data-gathering S/W, and the road sign detection S/W. In the second phase of the research, the research team developed an integrated program that recognizes the character and code regions from the road sign visual information, and established a database. Furthermore, the research team established a database on the property information of arrows, the names of places by direction, and the route numbers. In the third phase of the research, the research team compared and analyzed the content and location information from the road sign system database and the RRAP survey results, and developed a technology that verifies the conformity, suitability, and connectivity. Finally, the research team developed an interactive module that links the RRAP and the road sign system, and conducted a test run of the prototype at the test bed in Gangnam-gu, Seoul City and in Cheong Na District, Incheon from May to September 2015.
 
To promote the applicability of the research outcomes, the research team presented and exhibited the research outcomes in national and international conferences and events, including the 1st Road Safety, Management Technology Meeting, the ADB Transport Forum in the Philippines, and the 25th World Road Congress Seoul 2015, and explained the developed technologies to a delegation from China as well as to the directors and staff from Mongolia who visited South Korea.
 
 

 

modified
▲ Conceptual map of RRAP

 

 

 

Activation of national and international pattern recognition and visual information acquisition industry

 

The most important outcome of the research is the development of the state-of-the-art RRAP, which automatically detects and analyzes road signs. In particular, the automatic recognition technology of information in Korean from images taken with the national light, in convergence with the autonomous driving technology, is expected to pave the way for the advanced road information acquisition technology. Furthermore, the development of a multisensor-based road sign survey vehicle and related S/W enabled the acquisition and extraction of high-precision location-based road sign information through which the technical foundations on which the road sign database could be quickly and efficiently established and revised can be built. In fact, the prototype road sign survey vehicle developed in the research performed a complete inspection of all the signs on the seven national highways, and the acquired data on images, locations, and properties were used to update the road sign system database.
 
Another research outcome is the maximization of the efficiency of the DB establishment process through the automation of the DB comparison and analysis tasks for updating road sign information. It established an efficient and rapid road sign database system through which the availability and reliability of the road sign system can be improved, and that can offer high-quality public service. As the first research project for commercializing the extraction of road sign property information, its outcomes will be expanded and applied to the future research on establishing a DB for various road facilities. Its economic outcomes include saving KRW10 billion in data establishment costs and KRW5 billion in analysis costs by maximizing the DB establishment and updating tasks for the road sign system. Additionally, the research ensured safety in the work environments by reducing the survey staff and duration, and drastically saved on the total project costs by improving the accuracy of the extracted information.

 

 

RRAP2 RRAP3
▲ RRAP survey equipment ▲ Vehicle equipped with RRAP

 

 

It is expected that such outcomes from this study will be widely applicable to the private and public sectors. It is also expected that the research outcomes will continue to be used in the “Coordinate-based Road Sign Database Renewal Project” on national highways being promoted by the Ministry of Land, Infrastructure, and Transport based on Government 3.0, and that the introduction of the RRAP will be initiated by the Road Management Office for efficient road maintenance and repair.

Moreover, these outcomes will be widely used for establishing a traffic facilities database and visual information of various road facilities for Vworld, an open platform for 3D spatial information in South Korea being built by MOLIT, and for addressing the issues of numerous outdoor billboards and standing signboards. Finally, it is expected that through the development of road- and transport-related IT technologies, the research outcomes will activate the national and international pattern recognition and visual information acquisition industries.
 
 
 
List
top

프린트