The system detects that a child is left in the car
The system, researched and developed by a group of authors at the Industrial University of Ho Chi Minh City, can quickly and accurately detect the presence of people in the car when the vehicle is not in operation.
.jpg)
Testing the system in cars. Photo: NNC
In Vietnam and around the world, many incidents of forgetting children in cars have caused serious consequences.
When the car's engine is turned off and the doors are closed, the interior temperature will rise rapidly, even on days with an air temperature of about 21℃. Because a child's ability to regulate heat is not fully developed, this condition can lead to heat stroke and stroke. Children's safety should be ensured under adult supervision. However, while being careless, drivers and parents can leave children in the car.
Face with that fact, a group of authors at the Department of Dynamic Technology, Industrial University of Ho Chi Minh City, has researched and developed a system to warn people in the car, when the car is not in operation. This research applies artificial intelligence (AI), thermal sensors, and cameras to identify people and moving objects on the vehicle when the vehicle is stopped. The system uses the Python programming language, which controls the collection and processing of signals for the central processor.
The built-in system includes an IMX219-160IR Camera with integrated infrared sensor; Central processing unit (including Wireless NIC For Jetson Nano circuit, WiFi/Bluetooth and SIM7600G-H 4G/3G/2G/GNSS Module for Jetson Nano). The central processing unit recognizes using Machine Learning (machine learning, or a branch of AI), to classify people and objects in each image frame received from the camera. When the car is turned off, the doors and windows are closed, the system will automatically start within 1-2 minutes and identify people through the infrared sensor camera. Then, make an alarm and transmit the visual information to the smartphone.
The team tested the system in situations where it could actually happen such as Darkening the image (underground parking or in the evening); brighten the image (daylight); blur, image noise; create black boxes that obscure an infant's eyes, mouth, or nose. When the vehicle is turned off and closed, the system accurately identifies the person in the above hypothetical cases.
The system is tested in the Camry XLE (camera, processor are installed in the rear seats). When the car engine is off, the door is closed, the glass is closed with a person lying on the back seat of the car, the system gives a quick warning (about 5 minutes), precisely, by turning on the relay that activates the alarm, alarm lights. After the bell and warning light are activated, SIM7600G-H 4G/3G/2G/GNSS immediately sends image data to identify people in the car and sends a message to the user's phone.
The authors' research was reported at the 3rd Young Science Conference of the Industrial University of Ho Chi Minh City.
ctngoc