Commercial Display

Global

Global - English

Asia

Europe

Oceania

ANZ - English

Latin America

North America

Middle East and Africa

Full Analysis Server

The full analysis software is an image fusion analysis system that is able to realize the operation and flexible scheduling of multiple algorithms based on algorithm group. It can operate in various computing resource platforms to realize the cluster management and scheduling of computing resource.
Flexible Scheduling: Supports flexible scheduling of different intelligent algorithms like facial analysis, video analysis, and human body analysis in accordance with demands.Supports structured data extraction and modeling in accordance with demands.
Full Analysis: Adopts unified computing resource management and scheduling. Integrates different computing resources across multiple hardware platforms as one to play their max. effect.
Openness: Adopts an open structure to have unified management and scheduling of multiple algorithms to meet demands. Supports data openness to open video or picture data after structuralization via standard interface or protocol. Support development and application by the third party platform.

View Less

Data Sheet

DS-IF0100-AI

Specification
Resources
Accessories

Data Sheet

Face Analysis
Facial Video FormatResolution: 720 P to 8 MP Real-time video stream: H.264, Hikvision H.265
Video Frame Rate Requirement12 to 30 frames per second (25 frames per second recommended)
Face Picture FormatSupports format of PNG, TIF, BMP, JPEG
Face Picture RequirementFace analysis target: Resolution: 48 × 48 to 8000 × 8000 (64 MP) Size: ≤ 8 MB Pupil distance: 20 to 300 pixels recommended Face angle deviation: Face analysis: ≤ 35° (vertical); ≤55° (horizontal) Face picture of list library: ≤ 25° (vertical); ≤30° (horizontal)
Max. Number of Arming ListsT4: 3 million
Face Video Analysis and ComparisonEach T4: 30 channels (2 MP) 12 channels (2 MP to 5 MP) 6 channels (5 MP to 8 MP)
Face Picture Analysis and ComparisonEach T4: 170 pieces/second
Face AttributeSmile, glasses, mask

Vehicle Analysis
Data TypePicture
Picture FormatJPG/JPEG, BMP, PNG/TIFF, TIF, and static GIF
Vehicle Picture Requirement256 x 256 pixels to 9 MP Vehicle width: 100 to 300 pixels recommended License plate width: 80 to 120 pixels recommended Angle deviation: less than 30°(horizontal) and 25°(vertical)
Vehicle AttributesVehicle model, Vehicle color, license plate number, Vehicle main brand.
Analysis Performance of Vehicle Picture Each T4: 28 pieces/second (2 MP ) 28 pieces/second (5 MP ) 28 pieces/second (8 MP )

Video Structuralization
Data TypePicture and Video
Video FormatH.265;H.264
Video Frame Rate Requirement12 to 30 frames per second (25 frames per second recommended)
Picture FormatJPEG(JPG), TIF, BMP, PNG, partial GIF(static diagram)
Human Body Picture RequirementResolution: 256 × 256 to 8000 × 8000 (64 MP) Human body width: 70 to 300 pixels recommended Completeness: the blocking area should be less than 30 percent of the target Angle deviation: the depression angle should be no greater than 45°
Vehicle Picture RequirementResolution: 256 × 256 to 8000 ×8000 (64 MP) Vehicle width: 100 to 300 pixels recommended License plate width: 80 to 120 pixels recommended Angle deviation: less than 30°(horizontal) and 25°(vertical), modelling for both the front and rear
Analysis of Human Body and Vehicle VideosEach T4: 25 channels (2 MP) 16 channels (2 MP to 5 MP) 10 channels (5 MP to 8 MP)
Analysis of Human Body and Vehicle PicturesEach T4: 38 pieces/second (2 MP ) 33 pieces/second (5 MP ) 28 pieces/second (8 MP )
Human Body AttributeTarget direction/size/speed, coat/trousers type/color, hair style, riding, carrying bag, glasses, hat, mask, number of person on bicycle, etc.
Vehicle attributeTarget direction/size/speed, license plate number, vehicle color, license plate color, vehicle type, vehicle master brand, driver/front seat safety belt, etc.

Product Performance
Authorization TypeSoft Encryption
Protocol TypeISAPI

Product Description
System VersionCentOS 7.8
Device TypeSoftware only

Hardware Specifications
CPU1. Number of T4 cards: n*T4，n<=3 CPU requirements: Intel 6230(20-core, 2.1Ghz )*1 or above 2. Number of T4 cards: n*T4，3=<n<=8 CPU requirements: Intel 6230(20-core, 2.1Ghz )*2 or above
Memory1. Number of T4 cards: n*T4，n<=3 Memory requirements: 96G DDR4 2. Number of T4 cards: n*T4，3=<n<=8 Memory requirements: 96GB DDR4+32GB DDR4*（n-3）
Hard Disk480G SSD*1
Network InterfaceGigE x 2 or above
GPU CardNVIDIA ® Tesla ® T4 recommended

Face Analysis

Facial Video Format

Resolution: 720 P to 8 MP Real-time video stream: H.264, Hikvision H.265

Video Frame Rate Requirement

12 to 30 frames per second (25 frames per second recommended)

Face Picture Format

Supports format of PNG, TIF, BMP, JPEG

Face Picture Requirement

Face analysis target: Resolution: 48 × 48 to 8000 × 8000 (64 MP) Size: ≤ 8 MB Pupil distance: 20 to 300 pixels recommended Face angle deviation: Face analysis: ≤ 35° (vertical); ≤55° (horizontal) Face picture of list library: ≤ 25° (vertical); ≤30° (horizontal)

Max. Number of Arming Lists

T4: 3 million

Face Video Analysis and Comparison

Each T4: 30 channels (2 MP) 12 channels (2 MP to 5 MP) 6 channels (5 MP to 8 MP)

Face Picture Analysis and Comparison

Each T4: 170 pieces/second

Face Attribute

Smile, glasses, mask

Vehicle Analysis

Data Type

Picture

Picture Format

JPG/JPEG, BMP, PNG/TIFF, TIF, and static GIF

Vehicle Picture Requirement

256 x 256 pixels to 9 MP Vehicle width: 100 to 300 pixels recommended License plate width: 80 to 120 pixels recommended Angle deviation: less than 30°(horizontal) and 25°(vertical)

Vehicle Attributes

Vehicle model, Vehicle color, license plate number, Vehicle main brand.

Analysis Performance of Vehicle Picture

Each T4: 28 pieces/second (2 MP ) 28 pieces/second (5 MP ) 28 pieces/second (8 MP )

Video Structuralization

Data Type

Picture and Video

Video Format

H.265;H.264

Video Frame Rate Requirement

12 to 30 frames per second (25 frames per second recommended)

Picture Format

JPEG(JPG), TIF, BMP, PNG, partial GIF(static diagram)

Human Body Picture Requirement

Resolution: 256 × 256 to 8000 × 8000 (64 MP) Human body width: 70 to 300 pixels recommended Completeness: the blocking area should be less than 30 percent of the target Angle deviation: the depression angle should be no greater than 45°

Vehicle Picture Requirement

Resolution: 256 × 256 to 8000 ×8000 (64 MP) Vehicle width: 100 to 300 pixels recommended License plate width: 80 to 120 pixels recommended Angle deviation: less than 30°(horizontal) and 25°(vertical), modelling for both the front and rear

Analysis of Human Body and Vehicle Videos

Each T4: 25 channels (2 MP) 16 channels (2 MP to 5 MP) 10 channels (5 MP to 8 MP)

Analysis of Human Body and Vehicle Pictures

Each T4: 38 pieces/second (2 MP ) 33 pieces/second (5 MP ) 28 pieces/second (8 MP )

Human Body Attribute

Target direction/size/speed, coat/trousers type/color, hair style, riding, carrying bag, glasses, hat, mask, number of person on bicycle, etc.

Vehicle attribute

Target direction/size/speed, license plate number, vehicle color, license plate color, vehicle type, vehicle master brand, driver/front seat safety belt, etc.

Product Performance

Authorization Type

Soft Encryption

Protocol Type

ISAPI

Product Description

System Version

CentOS 7.8

Device Type

Software only

Hardware Specifications

CPU

1. Number of T4 cards: n*T4，n<=3 CPU requirements: Intel 6230(20-core, 2.1Ghz )*1 or above 2. Number of T4 cards: n*T4，3=<n<=8 CPU requirements: Intel 6230(20-core, 2.1Ghz )*2 or above

Memory

1. Number of T4 cards: n*T4，n<=3 Memory requirements: 96G DDR4 2. Number of T4 cards: n*T4，3=<n<=8 Memory requirements: 96GB DDR4+32GB DDR4*（n-3）

Hard Disk

480G SSD*1

Network Interface

GigE x 2 or above

GPU Card

NVIDIA ® Tesla ® T4 recommended

Face Analysis