Image Segmentation

Leveraging industry-leading image segmentation algorithms, this system pioneers embedded real-time segmentation of road hazard areas, accurately identifying danger zones and triggering efficient alerts based on warning strategies with low false alarm rates. For container space management, it enables real-time identification of vacant areas within cargo holds and automatically calculates space utilization rates. This capability facilitates optimized cargo layout planning, maximizes space efficiency, and ultimately reduces costs while enhancing operational productivity.

Character, Symbol Detection

Powered by intelligent vision technology, this system is widely applied across road and warehouse scenarios. In road applications, it supports recognition of multiple signs (speed limits, stop signed, etc.) and characters (e.g., license plates), enabling regulatory violation alerts, incident documentation and traffic supervision. In warehouse applications, it intelligently indentifies operational signs and triggers coordinated safety controls for forklifts (e.g., alarms/speed reduction) to enhance work safety.

Target Tracking

Based on a monocular camera,the automatic passenger counting function adopts deep learning technology to perform real-time identification and motion tracking of people in the images collected by the camera, so as to achieve high-accuracy passenger flow statistics. lt also has the functions of camera monitoring and video recording while carrying out passenger flow statistics, which not only save sequipment costs and installation costs, but also brings convenience to subsequent equipment operation and maintenance.

Detecting Abnormal Image

Employing deep learning models,our product accurately monitors camera operational status in real-time (Normal/Dirty/Occlusion/Glare), triggering tiered alerts for distinct anomalies to ensure monitoring quality. Detection is can be carried out in both visible light and infrared scenarios, enabling effective 7*24 surveillance.

Predicting Motion Status

The real-time image analysis enables precise anticipation of forklift movement states (lifting/lowering, advancing/reversing). It intelligently coordinates with equipment, which includes: Automatically activating laser beams during fork arm elevation to assist positioning, and deactivating lasers when forks are loaded or idle. Disabling reverse camera alarms during forward motion.This technology significantly boosts operational precision and efficiency,optimizes energy consumption, and ensures operational safety.

DMS Monitoring

Through multi-layer neural network technology, video image analysis technology, clustering and neural network algorithm, any abnormal driving behaviors (fatigue, distraction, smoking, calling, no mask) during driving can be accurately and real-time monitored, both visual and audio alarms will be actively offered to remind drivers to work properly and drive safely. This algorithm also support driver face detection and authentication, no driver status detection.

IMS Monitoring

Designed for commercial vehicles (buses/trucks), our system employs posture assessment and object detection technologies to conduct overhead monitoring of the entire cabin. It monitors driver and passenger behaviors, including cabin occupancy counting, driver status detection (distraction, hand-off-wheel gesture and unauthorized absence), and safety compliance detection (seatbelt fastening and smoking). Upon detecting violations, the system triggers tiered alerts to mitigate driving risks, ensuring passenger safety and regulatory compliance for fleet operations.

AI ISP

Based on an AI image enhancement engine, this feature intelligently optimizes images captured by the camera (Auto Exposure/Noise Reduction/White Balance/Color Restoration). It significantly improves image clarity and detail representation, especially in low-light environments, thereby enhancing target detection and recognition performance.

Image Deblurring

This AI algorithm processes blurry images captured by vehicle-mounted cameras during high-speed travel or on bumpy roads, effectively restoring critical image details. Currently it is primarily applied to license plate recognition.

Monocular Distance Measurement

Employing monocular ranging technology, this feature accurately calculates distances between the camera and targets (pedestrians/vehicles). It enables risk-tiered early warnings through fusion with sensors (e.g., radar), delivering core functionalities for OE ADAS systems like forward vehicle distance estimation, pedestrian detection, and collision warning.

Panorama Stitching

Using panoramic imaging and display technology, this feature generates real-time 360° seamless 1080P surround views. It continuously perceives vehicle's surrounding environment during operation, collects data for analytical processing, provides early warnings for potential hazards, enhances driving safety and operational efficiency.