Get demo
Get demo
Architecture
Solution
Scanning devices
variations
Description
Dimensions
Applications
Disadvantages
2D
Limited depth perception, only captures flat images
High resolution, network connectivity, flexible for industrial use
Matrix industrial IP camera, 2D
Description
Dimensions
Applications
Advantages
Disadvantages
1D
Continuous processes like sorting, scanning
High speed, suitable
for processes requiring fast data acquisition
Can only capture one line of data
at a time, no depth information
A 1D camera that captures line-by-line imaging, commonly used for conveyor belt scanning
Linear Camera
A 2D camera that captures images and video, often used in industrial automation
Surveillance,
quality control, process
monitoring
Advantages
Description
Dimensions
Applications
Disadvantages
3D
Limited range and resolution in large environments
Captures both visual and depth data, useful for 3D mapping and real-time tracking
RGBD Camera
Description
Dimensions
Applications
Advantages
Disadvantages
3D
Autonomous vehicles, terrain mapping
Extremely accurate distance measurement, works in various lighting conditions
Expensive, high power consumption, and limited in extreme weather
Uses laser light to measure distances, creating 3D models of environments
LiDAR
A 3D camera that captures both color (RGB) and depth information (D)
Object recognition, robotics, motion tracking
Advantages
Description
Dimensions
Applications
Disadvantages
2D
Lower resolution, cannot detect fine visual detai
Can see in the dark, identifies heat sources, non-contact temperature measurement
Thermal Imaging Camera
Description
Dimensions
Applications
Advantages
Disadvantages
1D
Document digitization, industrial inspection
Precise capture of fine details, useful for creating 3D models of objects
Limited to static objects, not suitable for real-time applications
Devices used to capture and digitize physical objects or documents
Scanning Devices
Captures infrared radiation to produce heat maps and detect temperature variations
Night vision, temperature monitoring, fire detection
Advantages
Software and hardware
specifics
Data Analysis Features
Primary Programming Language
Main Framework for Vision Model Development
Additional Tools
Various open-source Python libraries used for data analysis and neural network training
Vision Interfaces
Detection
Capture
Input:
Input:
Detection and classification for each image or segment of the video
Output:
Output:
Position and angle data for capturing actions
Images or video in RGB-D format
Images or video
Artificial Intelligence System
by True Detection
Developed by True Detection, this AI system excels in the recognition and analysis of recyclable materials by type and color. It operates on any clouds as well as the company’s internal servers.
The programming is done using Python, with the neural network analytics being proprietary software currently in the registration phase.
Unique Architecture
True Detection’s AI system leverages a custom neural network architecture with the following key features
Other state-of-the-art neural architecture innovations
Foundation
Structure
Advanced Techniques
Custom neural network based on CenterNet,
with a unique backbone
Multiple layers
of Convolutional Neural Networks (CNN)
of Convolutional Neural Networks (CNN)
Incorporates
cutting-edge methods:
Focusing on deep learning and advanced feature extraction
FPN (Feature Pyramid Networks)
Separable Convolutions
Deformable Convolutions
SE-Block
(Squeeze-and-Excitation Network)
External VS Internal capacities
Location of Data Processing
External Capacities
Data analysis takes place on remoted servers or cloud platforms
Data analysis happens locally on a facility servers at their premises
Internal Capacities
Data Storage
External Capacities
Information is stored externally on cloud servers managed by a third party
Data is stored on local servers owned and managed by the customer
Internal Capacities
Control Over Data
External Capacities
Limited control; the provider manages data storage and processing infrastructure
Full control over data storage and processing, including security protocols
Internal Capacities
Latency / Performance
External Capacities
Higher latency due to data being processed externally, which may cause delays
Low latency with faster performance, especially for real-time data processing
Internal Capacities
Scalability
External Capacities
Easily scalable; cloud providers can allocate more resources on-demand
Limited scalability based on the available on-premises hardware
Internal Capacities
Security
External Capacities
Security is managed by the third-party provider, but the data is subject to their policies
Full control over data security with customizable security measures
Internal Capacities
Costs
External Capacities
Lower upfront costs but ongoing subscription
or pay-per-use fees
or pay-per-use fees
Higher initial capital expenditure for infrastructure, with lower long-term operational costs
Internal Capacities
Data Privacy & Compliance
External Capacities
May face challenges with privacy regulations depending on the provider’s location and policies
Easier to comply with strict local data privacy regulations, as data stays on-premises
Internal Capacities
Reliability
External Capacities
Redundancy and backup systems are usually offered by cloud providers
Reliability depends on internal infrastructure and IT management practices
Internal Capacities
Internet Dependency
External Capacities
Requires a stable internet connection; outages affect data processing
Does not require an internet connection; operations continue regardless of internet availability
Internal Capacities
Customization
External Capacities
Limited customization in data processing; relies on the provider’s configurations
Full customization for specific processing needs and workflows
Internal Capacities
Reporting
Reporting System
Device and Network Notification System
APCS Integration
Process Indicator Reporting
Third-party Device Control
Incident Logging
Robotic Process Automation
Emergency signals
and automated notifications via SMS, email and control system
Ensuring complex signals are sent to equipment
or dispatchers
for emergency stops
or other critical actions
Real-time statistics, user-friendly dashboards, seamless integration into corporate databases, and CRM systems
We develop automation mechanisms, such as automatic removal of foreign objects or defective products from conveyors, and system-directed stream division
Automated creation
of incident reports using standardized templates
Installation
and integration of robots powered by our neural network, enhancing automation capabilities
Get demo
Сообщение об успешной отправке!
© 2024 True Detection
Technical Vision and Artificial Intelligence Laboratory
UAE, Abu Dhabi, Masdar City,
Incubator Building
Incubator Building