Favourite
IVS-PB8000-E-GU1

Intelligent Server for Public Event Detection

Key Points
    > Advanced structured + behavior analysis deep-learning algorithm with world-class target detection rate and behavior analysis accuracy
    > Dynamic loading of algorithm and chip separation to enhance system robustness.
    > Universal server with PCIE slot design that enables recycling old servers to reduce costs.
    > Based on video cloud architecture, it supports standalone and clustered deployment that satisfies various requirements for capacity expansion.
    > The server can be sold with a client or separately, and it can work with third-party devices.


Specifications

System

Processor

One Intel Xeon E3-1275 V5, 3.6 GHz, 4C/8T

Intelligent Analysis Card

One high-performance analysis card

Operating System

CentOS Linux release 7.4.1708 (Core)

Memory

Two 8 GB DDR4 memory, maximum 4 slots.

Disk

One 3.5”4T disk, maximum 4 disks.

7.2K RPM SATA 6 Gbps 512n 3.5 inch

Public Event Detection

Multi-rules Application

Multiple rules can be configured and run at the same time.

Real-time Display

Supports displaying rule detection zone and target detection box in real-time video.

Rule detection zone and target detection box will flash when an alarm is triggered.

Crowd Density Detection

Supports to set max 8 monitoring areas in one video channel; give an alarm when number of people in the area exceeds the threshold for certain period; give an alarm when area density exceeds the threshold for certain period; support real-time statistics of total number of people in the video of output channel; support real-time statistics and output of number of people in each monitoring area.

Crowd density: Low density scene: People counting accuracy ≥80% (not required if there are less than 10 people);

High density scene: People counting accuracy ≥75% (20–50 people), people counting accuracy ≥80% (50+ people) (not required if there are less than 10 people).

People Counting

Supports real-time analysis on number of people entering and leaving one area in video channel;

Target pixel ≥60×60, and accuracy >93% (in the day, no obvious obstacles).

Tripwire

Automatically detects tripwire crossing;

Target pixel ≥60×60, recall rate ≥90%, and effective rate ≥90%.

Intrusion

Monitors the forbidden area and give an alarm;

Target pixel ≥60×60, recall rate ≥90%, and effective rate ≥90%.

Object Detection

Monitors suspected abandoned object in key protection areas, and give an alarm (bags, luggage and boxes);

Target pixel ≥60×60.

Loitering

Monitors people staying in key protection areas and give an alarm;

Target pixel ≥60×60, recall rate ≥90%, and effective rate ≥90% (less than 5 people).

Fast Moving

Detects fast moving and give an alarm;

Target pixel ≥60×60, recall rate ≥80%, and effective rate ≥80% (less than 5 people).

Fighting

Detects fighting and give an alarm;

The fighting target takes up 1/4 × 1/4 of the image.

Wrong-way Walking

Monitors pedestrians who walk in the wrong way in unidirectional areas, and give an alarm;

Target pixel ≥60×60, recall rate ≥90%, and effective rate ≥90% (less than 5 people, no obvious obstacles).

Escalator Anamoly

Supports real-time monitoring of sudden stop and reverse operation of escalators;

Target pixel ≥60×60.

Report Generation and Export

Exports alarm information in Excel, including channel name, event time, event name, event type, rule name, rule line, scene and target box.

Alarm Search

Search for alarm information by device, channel, type, time, and more.

Public Event Detection Performance

Resolution

720p and above.

Analysis Capability

People counting: One server can access and analyze video data of up to 32 channels of 1080p cameras at the same time.

Crowd density: One server can access and analyze video data of up to 32 channels of 1080p cameras at the same time (For cameras installed 9 m or more above the ground, up to 16 channels are supported).

IVS: One server can access and analyze video data of up to 32 channels of 1080p cameras at the same time (up to 8 cameras for fighting/escalator anomaly).

Public Event Detection Application

Scenario Type

Extensively applicable to airports, subway and railway stations, squares, office perimeters, places of business, supermarkets, and more.

Camera Installation Method

Front installation (recommended) and side installation.

Camera Installation Height

3–6 m is recommended; for crowd density analysis, 9–20 m is recommended.

Camera Installation Angle

10–75° for pitch angle is recommended, and target pixel should be ≥60×60.

Resolution

720p and above.

Port

Network Port

Two 1000 Mbps Ethernet ports

USB

2 front USB2.0 ports, 2 rear USB3.0 ports and 2 rear USB2.0 ports

VGA

2

Others

1 DVI port and 2 DP ports

General

Power Supply

100–240 V, 50–60 Hz, 6 A–3 A

Power redundancy

Single power supply

Power Consumption

≤ 400W

Operating Temperature

+10 °C to +35 °C (+50 °F to +95 °F)

Operating Humidity

10%–80% (RH)

Storage Temperature

+10 °C to +65 °C (+50 °C to +149 °C)

Storage Humidity

5%–95% (RH)

Gross Weight

16.00 kg (35.27 lb)

Net Weight

8.50 kg (18.74 lb)

Product Dimensions

43.50 mm x 438.50 mm x 550.00 mm (17.13" x 172.64" x 216.54") (H × W × D)

Packaging Dimensions

271.00 mm x 625.00 mm x 895.00 mm (106.69" x 246.06" x 352.36") (H × W × D)

Installation

Standard 19" rack installation with guide rail

Type

Model

Description

1U Intelligent Server for Public Event Detection

DHI-IVS-PB8000-E-GU1

1U Intelligent Server for Public Event Detection

* Limited to a maximum of 4 products
Reset
Compare