According to public survey data, one person per day in the world is injured by a car every seven minutes. Every two hours or so, a pedestrian is killed in a car accident. Especially in the national provincial roads, county roads, village intersections, non-light control intersections where drivers have poor visibility, pedestrians and non-motor vehicles cross the road, causing a large proportion of traffic accidents, and often a major casualty accident. Most of them have always lacked effective scientific and technological means to prevent such accidents.
GMI's pedestrian crossing warning system combines video detection, behavior recognition, IoT communication, low-power LED and other technologies. It uses artificial intelligence and machine vision technology to automatically recognize pedestrian behavior and judge pedestrians who are about to cross the road. Flashing spikes, warning screen warnings, etc., reminding the past car deceleration and ritual, guiding pedestrians to pass safely, thus effectively avoiding traffic accidents.
Product Networking Architecture
The pedestrian crossing early warning system consists of pedestrian detector, flash nail lamp and pedestrian crossing warning sign. When the pedestrian crosses the intersection and enters the detection area, the pedestrian detector can trigger the flashing road nails to shine, and at the same time, the signal can be transmitted to the electronic pedestrian crossing warning signs on both sides of the road to give courtesy to pedestrians, which can warn the vehicles to slow down and pay attention to the pedestrians. It is conducive to increasing the safety of pedestrians at the intersection and reducing traffic accidents.
Typical Application Scenarios
This product can be widely used in the field of intelligent transportation:
● No pedestrian crossing signal control section in the city;
● Crossing of factories, enterprises, and schools, and frequent intersections;
● The city, county, district, provincial road, county and township road sections are selected and intersected at the village;
● Industrial parks, industrial parks and internal roads.
Cooperating with traffic management platform can obviously improve the efficiency of road traffic management and reduce the probability of traffic accidents.
● The front-end detection equipment of the system adopts the machine learning method of motion region detection and HOG+SVM feature extraction, and detects the texture features of the pedestrian's head and shoulders in real time. When the extracted feature data is characterized as “Pedestrian”, the confidence is relatively high. An early warning signal can be generated. When the machine learning algorithm is training, the system introduces negative samples of the interference factors such as the marked lights, street lights, tree shadows, vehicles, rain and snow, so that the training algorithm has very strong scene adaptability and can be more effective. Pedestrian detection is performed using visual analysis.
● The system uses the LED lights and flashing stud lights on the crosswalk sign, which is far away from the distance, and the dynamic warning effect is better, which maximizes the driver's attention visually.