基于深度学习和自相似的物流安全场景创建早期火灾检测教学演示平台的研究

Early Fire Detection Teaching Demonstration Platform About Logistics Security Scenario Based on Deep Learning and Self-Similarity

火灾是一种具有很强破坏性和多发性的灾害之一,严重危害着物流过程中人们的生命财产安全。火灾发生早期多以烟雾和细小火焰为主要特征,而现有视频火灾检测算法极易忽略小目标特征,容易导致早期火灾发生误检漏检现象。结合早期火焰细小以及火焰和烟雾的自相似性特点,文章设计了一套基于深度学习结合火焰自相似性的物流安全领域早期火灾检测教学演示平台。该平台采用多线程处理,利用深度学习结合传统背景建模和目标检测算法对早期火灾进行检测记录,方便教学演示。该平台使用Python编程语言进行设计,学生通过简单学习便可在线操作演示平台进行实验,能够有效提升学生实际动手能力。

Fire is one of the highly destructive and frequent disasters,seriously endangering the safety of people's lives and property during the logistics process.In the early stages of fires,smoke and small flames are the main features,and existing video fire detection algorithms are prone to ignoring small target features,which can easily lead to false detection and missed detection in early fires.Combined with the characteristics of small early flame and self-similarity of flame and smoke,this paper designed a set of early fire detection teaching demonstration platform in the field of logistics safety based on deep learning combined with flame self-similarity.This platform adopts multi-threaded processing and utilizes deep learning combined with traditional background modeling and object detection algorithms to detect and record early fires,facilitating teaching demonstrations.The platform is designed using the Python programming language,and students can operate the demonstration platform online for experiments through simple learning,which can effectively improve their practical skills.