RMG Visual Anti-Collision System

It is an Intelligent active safety protection system custom-developed for large Rail-mounted Gantry Cranes (RMGs) in railway container stack yards. The system leverages multi-sensor fusion perception and artificial intelligence visual analysis technologies to achieve real-time detection, classification, and recognition of various obstacles within the RMG’s operating area, as well as prediction of their motion trends and intelligent collision-avoidance responses, thereby effectively ensuring operational safety and continuous equipment performance in railway stack yards.

Vision-first recognition and classification

Wide-angle HD industrial cameras deployed at the RMG’s door legs and saddle beams, leveraging a self-developed lightweight deep neural network model, are the first to perform real-time semantic segmentation and object detection within a 15-meter range, swiftly identifying obstacle categories (such as personnel, straddle carriers, and adjacent RMGs) and making preliminary assessments of their motion intentions.

Lidar-assisted

By integrating LiDAR point cloud data, high-precision 3D modeling and verification of obstacle contours and distances are achieved.

Centimeter-Level Electronic Fence and Adaptive Protection

Leveraging laser–vision fusion perception, a centimeter-level precision virtual electronic fence is delineated along the RMG’s operating path; it supports flexible configuration across multiple zones and angles; the fence status is updated in real time and can dynamically adjust the protection zone based on the operation plan and scene.

Intelligent Deceleration and Smooth Control

The system automatically adjusts the RMG trolley’s travel speed based on the collision risk level, enabling a smooth transition from “early warning” to “speed reduction” and then to “stop.”