|RESULTS OF FIELD TEST OF AUTOMATED MOTORCYCLE-RELATED ENFORCEMENT SYSTEM (UNITRAFFIC)|
Over the past 10 years, Korea's delivery sector has grown at an incredible rate, especially the number of food deliveries, which are mostly made by motorcycles. According to the National Statistical Office, the amount of online food service transactions from 2.96 trillion won in 2017 increased six times to 17.38 trillion won in 2020. (donga.com).
|Korean Government Efforts to Minimize Motorcycle Accidents. Research and Development of Automated Motorcycle-Related Traffic Law Enforcement System|
With the increase in delivery volumes, the number of motorcycles on the roads has also skyrocketed, as well as the number of traffic violations, which not rarely end in accidents. In this regard, the Korean government decided to pay more attention to this problem and find a way out of this situation, thus it instructed the Road Traffic Authority to study the issue.
So far, the most common and effective way to reduce accidents is fines that help drivers develop the habit of observing traffic rules. To do so, the violation must be detected properly, and the license plate of the violator vehicle must be recognized. Unlike four-wheeled vehicles, which have license plates both in front and rear, motorcycles have license plates only in the rear. And since, automated enforcement systems are shooting from the front, violating motorcyclists cannot be identified. Another problem in identifying violating motorcyclists is motorcycle license plate sizes and appearance, which differ from those recognized by most ALPR systems.
UNISEM, as a company that has experience in delivering systems detecting traffic violations by motorcyclists in Vietnam and Indonesia, has been advising the Road Traffic Authority in this matter.
After research conducted by the Road Traffic Authority, it was decided to allow execution of field test to several companies, which were pre-selected based on their compliance with the Road Traffic Authority requirements.
|A Field Experiment with UniTraffic for the Road Traffic Authority’s Research|
Among selected companies there was UNISEM. From December of 2020 to January of 2021, our team conducted a field test of the automated motorcycle-related traffic law enforcement system. In a designated area in Seongnam (Gyeonggi-do), equipment, including two cameras, one for license plate detection and one for traffic light detection, radar, led lights for night shooting, and a server with our software, UniTraffic’s VAS (Video Analytics System) were installed to prove the functionality of the UNISEM's solution.
|As was expected, UNISEM was the first company who completed the field test and submitted results. At the meeting on January 26 with Prime Minister's Office and Road Traffic Authority representatives, UNISEM’s team presented a report about work that was done and the performance of the UniTraffic’s VAS and LPDR in the Korean environment.|
|Two-wheeled Vehicle Traffic Analysis|
The detailed information cannot be provided but we can share some interesting statistical information regarding motorcycles from the observed area. The two figures below show how the number of motorcycles increased during lunch and dinner on both Sunday and Monday. But there is a slight shift in the peaks: on Sunday, the first peak(lunchtime) is at 1 pm and second(dinnertime), the highest, is at 6 pm, while on Monday the lunch peak is at noon, and the peak for dinner is at 8 pm. This difference appears to be due to the fact that Monday is a workday. The lunchtime in Korea is mostly from 11:40 am to 1 pm, and supposing people get to their home after 7 pm most of the food deliveries occurs at 12 pm and 8 pm which reflected in the graph.
|Figure 1, 2. NUMBER OF TWO-WHEELED VEHICLES PASSED OBSERVED AREA BY TIME.|
(December 6, Sunday - left, December 7, Monday - right)
|Figure 3 based on data also acquired from UniTraffic shows the trend of the motorcycle volume by weekday, how it changes through the week and find the highest and lowest points of the week. We are sure this kind of data can be particularly useful in the process of analyzing traffic and planning developments. |
|Figure 3. NUMBER OF TWO-WHEELED VEHICLES PASSED OBSERVED AREA BY WEEKDAY.|
|Motorcycle-related Violation Detection and License Plate Recognition|
And the most important function in the field test was data provision about violations. According to only one day data generated from the observation, the number of motorcycle related violations is over five hundred cases (December 3: speed violation 282, traffic light violation 149, non-use of helmet violation 76) which means 4 out of 10(December 3: total number of motorcycles 1,242) motorcyclists violate the traffic law.
After the violation is detected, UniTraffic's LDPR recognizes the violator's license plate. It was the first time UniTraffic operated in an unfamiliar environment of Korean roads, but the recognition rate of license plates was over 98.4%, excluding partial recognition and recognition failure due to covering, pollution, or damage. The cause of partial recognition was the difference in license plate by region, which could be improved by additional learning (Deep Learning), and the problem of covering and pollution require driver edification and regulation tightening up.
(Plan) 2nd Year Research
This year the Road Traffic Authority is planning to continue and conduct second year research and to issue equipment specifications standards as well as measures to improve the performance of automated motorcycle enforcement equipment.
An increase in motorcycle accidents is a result of the fact that the number of motorcycles in Korea is increasing and many motorcyclists do not comply with traffic laws. We believe, UniTraffic with its excellent violation detection and license plate recognition functions might become an effective tool in the process of enforcing motorcyclists to comply with traffic regulations.
UNISEM stock price has an increase of 3.70% as of February 9th, and the strength of the contract remained strong at 500%.
UNISEM announced on January 26th that its operating profit based on its consolidated financial statements last year increased 36.6% from the previous year to KRW 32.8 billion through the disclosure of a change of 30% or more in sales or profit and loss structure. Sales also increased by 20.1% from the previous year to 215 billion won.