How It Works — True Detection

Our Technology: Neural Networks, Thermal Imaging, and Intelligent Analysis

The True Detection platform receives data from cameras and scanners, analyzes it with neural networks, and delivers ready-to-use results to your business systems.

True Detection technologies

How the Technology Works

1
Video Stream Devices feed data into the system
2
Frame Processing Normalization, filtering, and preparation for analysis
3
Neural Analysis Real-time GPU/CPU inference
4
Recognition Objects, attributes, and events in the frame
5
Event Logging The system records incidents and metrics
6
Analytics Data aggregation, dashboards, and KPIs
API
Notifications
Reports

Supported Devices

The system works with a wide range of data sources — from standard cameras to thermal imagers and LiDARs.

Line-Scan Camera (1D)

Captures the image line by line as the object moves. Ideal for conveyors and product quality control.

Area-Scan IP Camera (2D)

Standard video surveillance for object and event analysis: people, vehicles, perimeter, and indoor spaces.

RGBD Camera (3D)

Adds scene depth to the image. Precise object geometry, volume measurement, and 3D analytics.

LiDAR

Laser scanning of the environment. Precise distances and object contours even in low light.

Thermal Imaging Camera

Operates in the infrared range. Sees objects in darkness, smoke, and fog; monitors overheating.

Scanning Devices

Specialized solutions for scanning and measuring objects: dimensions, defects, and marking.

Artificial Intelligence System

Object Detection

People, vehicles, and products in the frame — in real time.

Tracking

Following objects across cameras and over time.

Product Quality Control

Automatic inspection of products for defects and conformity to standards.

Segmentation

Object boundaries with pixel-level precision.

Event Classification

The system distinguishes normal situations from anomalies.

Attribute Recognition

Color, type, dimensions, and other object characteristics.

Core Platform Components

Modular architecture: each component operates independently and scales separately.

AI Models

Trained neural networks for detection, recognition, and segmentation.

Detectors

Object detection modules tailored to the customer's tasks.

Tracking

Object tracking and trajectory reconstruction.

Analytics

Event aggregation, process metrics, and dashboards.

API

Open interface for integration with external systems.

Data Storage

Archive of events, frames, and metadata with fast search.

Performance

Real Time Minimal latency from frame to event (~1 sec)
GPU Support Parallel processing on graphics accelerators
CPU Support Runs on CPUs without dedicated GPUs
Scalability Hundreds of cameras in a single cluster
Multi-Threading Simultaneous processing of multiple streams

Deployment Options

Your Infrastructure

On-Premise

Installed on the customer's servers. Data never leaves the facility.

  • Full control over data and storage
  • Maximum confidentiality and security
  • Works without an internet connection
  • Deep customization to enterprise processes
Our Cloud

Cloud

Processing in our cloud. Fast start without your own hardware.

  • Launch in days, not months
  • Elastic scaling to match the load
  • Updates and maintenance on our side
  • Convenient for distributed facilities

Integrations

Open protocols and interfaces — the platform integrates seamlessly into your existing IT infrastructure.

REST API
RTSP
ONVIF
MQTT
Kafka

Reporting & Automation

Alerts & Notifications

  • Alert system and emergency notifications to messengers
  • SMS and e-mail notifications
  • Information dashboards

Reporting

  • Process performance reporting
  • Automatic report generation
  • Incident logging

Integration

  • Integration with SCADA / industrial control systems
  • Connection to corporate databases
  • CRM integration

Automation & Control

  • Control of external devices
  • Control of robotic systems
  • Automation of sorting and production processes