The utilization of haul trucks within the mining industry presents a myriad of safety challenges, primarily due to their considerable size. Haul trucks serve as indispensable assets in large-scale construction projects, owing to their unparalleled capacity for transporting heavy loads. Tailored specifically for such rigorous tasks, these trucks are restricted to traversing non-crowded roads to uphold safety standards. Nonetheless, their sheer size and design contribute to numerous blind spots, rendering them challenging to operate. In this study, we address the issue of blind spots in haul trucks by proposing a cutting-edge machine vision system. Our innovative approach integrates machine vision technology with advanced Deep Neural Network (DNN) algorithms to detect objects within the truck’s blind spots and promptly alert the driver in real time. Leveraging a comprehensive dataset of images, the DNN is trained to identify objects of varying sizes and shapes accurately. To facilitate seamless monitoring, our system incorporates a visual display unit installed in the truck’s cab, offering the driver a real-time panoramic view of the surroundings. Initial testing conducted in controlled laboratory settings yielded promising results, with the DNN algorithm demonstrating high accuracy in object detection. To validate our concept, we opted to employ a standard car as a surrogate for the haul truck. Equipped with four strategically positioned cameras providing a 360 -degree field of view, we scaled down the truck’s dimensions to match those of a car. Utilizing the D455 Intel RealSense camera, we precisely measure the distances of detected objects and issue alerts to the driver within a 4 -meter range via a built-in buzzer. Our system is further fortified with the inclusion of the Jetson Orin Nano developer kit. Real-world testing conducted in a moving vehicle showcased the effectiveness of our system, garnering favorable results. Our endeavor underscores a significant step towards enhancing safety and operational efficiency in haul truck operations through innovative machine vision solutions.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Advancing Haul Truck Safety: A Deep Neural Network-Based Driver Assistance System for Blind Spot Detection


    Contributors:

    Published in:

    Publication date :

    2024-12-02


    Size :

    4698226 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English




    DRIVER ASSISTANCE SYSTEM FOR REDUCING BLIND-SPOT-DETECTION FALSE ALERTS

    SCHWINDT OLIVER / GARCIA BORDES MARIA EUGENIA | European Patent Office | 2018

    Free access

    Driver assistance system and method for providing blind spot image

    LEE HO YEOL / LIM DONG HWAN | European Patent Office | 2021

    Free access


    BLIND SPOT ASSISTANCE DEVICE

    OBATA MASATO / TAKAHASHI YUICHI / KASAHARA TSUYOSHI | European Patent Office | 2017

    Free access