Welcome to the official website of Shanghai Qnode Technology Co., Ltd. !

Active Obstacle Detection System

Active Obstacle Detection System

The train active obstacle detection system is an active safety warning system that prevents accidents in advance against obstacles such as people, luggage, iron shoes, etc. on the track, and detects obstacles within the travel distance of the train. The current mainstream sensors are used to collect video within a certain range in front of the train, and then the video stream is transmitted to the host through communication between the sensor and the host. After the host receives the video stream, it processes the video algorithm to determine whether there are obstacles in the front range and the type of obstacles, and then transmits it to the vehicle, outputs an alarm signal, and provides intuitive early warning prompts, thereby effectively preventing accidents....
Product Content
Technical Parameters

System Introduction

The train active obstacle detection system is an active safety warning system that prevents accidents in advance against obstacles such as people, luggage, iron shoes, etc. on the track, and detects obstacles within the travel distance of the train. The current mainstream sensors are used to collect video within a certain range in front of the train, and then the video stream is transmitted to the host through communication between the sensor and the host. After the host receives the video stream, it processes the video algorithm to determine whether there are obstacles in the front range and the type of obstacles, and then transmits it to the vehicle, outputs an alarm signal, and provides intuitive early warning prompts, thereby effectively preventing accidents.

Active Obstacle Detection System(图1)

System Features

1. Sensor acquisition and transmission


Using the characteristics of high resolution and high transmission speed of network cameras, the camera transmits the collected data to the host at a high rate, maximizing performance and reducing time waste.


2. Host deep learning algorithm processing


The host receives the collected video and processes it by framing, predicts each frame of the image stream through a deep learning neural network, and outputs its (region of interest) and the current prediction percentage (acc) through the network model attention mechanism and the improvement of the secondary algorithm.


3. Safety prompt warning


When the image and evaluation indicators output after algorithm processing are thresholded, the effect of reducing false alarms is achieved. When the evaluation indicators exceed the original set threshold, the signal is transmitted and the warning prompt is given.


4. Data recording


After the sensor is turned on, each video stream transmitted to the host algorithm module is thresholded and the output result after algorithm processing is saved locally so that potential safety hazards and problem reproduction can be better judged.

System Features

1. Safety certification


The system has passed the SIL2 safety level certification and has been verified in multiple operating projects.


2. High reliability


The system is mainly processed by software-side algorithms, with fewer hardware accidents, low failure rate, low replacement rate, and can conduct effective secondary reviews and software upgrades and optimizations.


3. Easy to use


The system is processed on the software side, without the complexity of hardware, and can more intuitively and effectively show the system processing effect.