Time Eye – AI Traffic Monitoring

V2X-Mobile Edge Computing Unlock New Possibilities

A smart city is a technologically modern urban area that uses different types of electronic methods, voice activation methods and sensors to collect specific data. In public safety section, V2X-mobile edge computing unlocks new possibilities to provide more efficiency ways to manage traffic and collect data. Sensors are key components of an intelligent device, on-vehicle sensors capture dynamic scene on street including traffic signs, road infrastructure, and traffic conditions. From device to edge, from edge to cloud, pairing with 5G, AIoT has numerous potential to expand business usage by its low latency and greater speed in the transmission, making sure to achieve data agility, mobility, and connectivity.

5G Fueling the Growth of AIoT

The rapid changes by technology, industries and societal patterns and process, all of these give the credit to increasing interconnectivity and smart automation. With 5G low-latency and high speed characteristics, the AI-based applications could react nearly real-time to the data that generated by 5G network. Therefore, 5G plus AIoT enables the digital transformation in transportation, smart city, and smart factory, etc.
From device to edge, and from edge to cloud, it requires high speed and reliable network to make sure the data can be delivered in real-time, the technology that depends on artificial intelligence and machine learning system needs to be determined in a fraction of a second. Pairing with 5G, AIoT has numerous potential to expand business usage by its low latency and greater speed in the transmission, making sure to achieve data agility, mobility, and connectivity.

Scope of the Edge and Cloud Roadmap

AIoT Data Collection

Smart Traffic

Road infrastructure law enforcement in parking zone includes license plate recognition, pedestrian tracking, and passenger counting. AI computer can also detect road infrastructure and provide information to avoid dangerous situation such as real-time traffic information, potential hazard zone identification, adaptive signal coordination, heat map, and obstacle detection. As for public safety, the parking zone enforcement, vehicle/pedestrian tracking, security monitoring, accident prevention and license plate recognition are also the key to create safety transportation.

V2X-Mobile Edge Computing Unlock New Possibilities

1. Video Preprocessing

1-1 De-noise

1-2 Rolling shutter video correction

1-3 Photometric correction

2. Vehicle detection and Plate Recognition

2-1 Keras+Openvino

2-2 Model : Single Shot MultiBox Detector(SSD)

3. The recognition of bus parking area

3-1 Keras + OpenVINO

3-2 Model : Deep residual network (ResNet)

4. Vehicle tracking

4-1 OpenVINO

AIot Ready In Vehicle Edge Computing

Data Capture>>	Data Analysis>>	Data Category>>	Data Application
Smart Sensors	Edge Computing	AI Framework OpenVINO	Cloud Based Computing

BUS-1

BUS-2

BUS-3

Next Step

•Blood Pressure

•Talking & Driving

•Eating & Driving

•Fall in Sleep

•Fasten Seat Belt

•Over-Speeding

•Fare evasion

•Harassment

•Walking on the Bus

•Fall Analysis

H/W Architecture Description

The Edge AI vehicle computer combined with AI image recognition, real-time interpretation of the occupancy behavior of bus parking. The violation of videos and license plate recognition results can be returned to the AI Cloud image recognition cloud management system for storage.

Award