基于热成像和可见光的机器人自动寻火装置研发

RESEARCH AND DEVELOPMENT OF AUTOMATIC FIRE-SEEKING DEVICE BASED ON THERMAL IMAGING AND VISIBLE LIGHT FOR ROBOTS

  • 摘要: 为了解决消防机器人在实现自动灭火任务时不能智能进行火点识别和定位的现状,提出一种基于热成像和可见光的信息融合深度学习识别方法。进一步提出一种基于空间变换的火点定位算法,该算法能够给出准确的火点坐标信息。利用所提出的识别和定位方法,研发用于特种机器人的自动寻火装置。该装置采用NVIDIA Jetson Xavier NX工控机作为上位机,STM32为下位机,对火焰进行快速识别和火点定位,从而使机器人能智能实现灭火任务。实验结果验证了寻火装置在火焰识别和火点定位上的有效性。

     

    Abstract: To solve the current situation that fire-fighting robot cannot intelligently identify and locate fire points when it realizes automatic fire-fighting tasks, this paper proposes an information fusion deep learning recognition method based on thermal imaging and visible light. A localization algorithm was proposed for fire based on spatial transformation. The algorithm could provide the precise coordinate information of the fire. Using the proposed identification and positioning methods, an automatic fire-seeking device for a special robot was developed. The fire-seeking device used the industrial control computer NVIDIA Jetson Xavier NX as the controller of the upper computer to identify the flame and the STM32 as the controller of the lower computer. The fire-seeking device could quickly identify and locate the fire, so that the robot could perform the task of automatic fire extinguishing. The experimental results verify the effectiveness of the device in flame recognition and fire point positioning.

     

/

返回文章
返回