Smart Traffic - EDGEMATRIX Inc.

HOME
Edge AI Station α
Smart Traffic

Edge AI Station α AI Application

｜App Features｜

Draw lines and count the number of vehicles passing by vehicle type.

Counts entries from both directions of the line

External system integration using Node-RED with desired protocols and formats

｜App Intro Video｜

Applied to Smart Traffic (v1.0.0).(※Under development image.)

｜Usage Scenes｜

Gate passage management for factories and facilities

Cameras are installed at the entrances and exits of factories and facilities to automatically count vehicles passing through designated lines. This can be used to keep track of the number of trucks entering and leaving the facility and to analyze logistics volume by time period, eliminating the need for manual recording and human verification. By identifying and recording the IN/OUT direction, it also supports the visualization of traffic history.

Construction

Logistics Center

Parking Lot

Traffic volume survey and traffic flow analysis

Using camera footage installed at designated locations, vehicles passing through are automatically counted by vehicle type and the direction. This enables the collection of quantitative traffic volume data without manual labor, and can be used as basic data for understanding traffic congestion trends, analyzing traffic flow by time of day, and formulating road maintenance plans and urban transportation policies. Compared to conventional manual surveys and simple sensors, installation work is reduced and installation locations can be selected flexibly.

Intersections

Roads

Highways

School Zones

*Node-RED is a visual programming development tool for interconnecting hardware devices, APIs, and online services.

｜App Overview｜

Counts the number of vehicles passing through the line in real time and implements the following three functions.

Record the number of IN/OUT
passages by vehicle type and direction.

Setting the timing for clearing
the cumulative count value

Visual understanding of the situation
through superimposed display

With Node-RED, a standard feature of AI Station, inference results from
AI applications can be processed and sent formatted for integration with external systems.

Generic Node-RED Flow (Post-Processing) Overview

Mail alert (sending)

Generic flow that determines intrusion into the ROI from the AI app's inference results and sends an email with a snapshot image.

*Customer mail server and mail account are required.

Since many AI Station use cases are expected to be deployed in on-premise environments or closed networks with no internet connection, the provision of web apps for visualization is optional for AI Station compatible apps (AIP), unlike the existing store apps (EAP). However, it is not necessary to provide a web app for visualization in an environment with internet access. However, for customers who install AI Station in an environment with internet access, we can provide a web app for visualization by sending the inference results to your website through post-processing by Node-RED. By preparing a web application for visualization, you can appeal your AI application to a wider range of customers, so we hope you will consider developing and providing a web application for visualization.

Grafana

*This screen is taken from the analysis screen of Kaleidoscope (a multifunctional people detection app).
We plan to use a general-purpose visualization system that can be built with OSS, consisting of Grafana + InfluxDB.
*We plan to provide a manual on how to build a general-purpose visualization system and a general-purpose Node-RED flow for sending visualization data to InfluxDB.
*Compliant with Grafana dashboard functionality.
*The number of vehicles by vehicle type and entry direction can be visualized in real time using graphs, tables, etc. The raw data used for visualization can also be downloaded in CSV format.

｜Function Details｜

Detection Methods

It has an automatic counting function for the number of vehicles passing through the line. It automatically counts the number of times an object crosses the line set by the user within the camera image. It's possible to identify the direction of entry (IN) and exit (OUT) and count them separately. Classification and aggregation by vehicle type (target class) is possible. Any combination of the following five classes can be set: “Car,” “Truck,” “Bus,” “Motorbike,” and “Bicycle.” The number of vehicles crossing each class is counted and overlaid in real time on the video. Overlay display function. Real-time cumulative passage counts for each target class, bounding boxes (detection target frames), line positions, and passage directions are overlaid on the video. Operators and facility managers can immediately grasp the on-site situation.

App Provider Company

Company Name:EDGEMATRIX Inc.

About Us:We contribute to society through the development of “Video Edge AI” infrastructure technology. Our mission is to implement “Video Edge AI” in society.

URL:https://edgematrix.com/en

Edge AI Station brochure
Download PDF ▶

Implementing "Video Edge AI" into various fields.