实验室火灾危险源高精度辨识方法优化仿真  

Optimization and Simulation of High Precision Identification Method for Laboratory Fire Hazard Sources

作　　者：徐永华 潘玉奇[2] XU Yong-hua;PAN Yu-qi(School of Computer Engineering,Jinling Institute of Technology,Jiangsu Nanjing 211169,China;School of Information Science and Engineering,Jinan University,Shandong Jinan 250022,China)

机构地区：[1]金陵科技学院计算机工程学院,江苏南京211169 [2]济南大学信息科学与工程学院,山东济南250022

出　　处：《计算机仿真》2022年第9期243-246,354,共5页Computer Simulation

基　　金：江苏省教育部产学合作协同育人项目(202101225010)。

摘　　要：针对目前方法对火灾危险源进行辨识过程中,由于未能在火灾危险源辨识前对数据进行平滑去噪处理,导致该方法在危险源辨识时,存在辨识精度低、最大热释放率与实际值相差大以及危险源增长系数高的问题,提出基于机器视觉的实验室火灾危险源辨识方法。方法首先对实验室数据进行全方位采集,通过对采集数据离群点滤除以及法向量的估计实现数据的平滑去噪;再使用差分进化法对机器视觉中的在线序列极限学习机进行优化;最后将平滑去噪后的实验室数据放入优化后的学习机中进行训练,从而实现实验室火灾危险源的辨识。实验结果表明,运用该方法识别危险源时,辨识精度高,最大热释放率与实际值相接近,危险源增长系数低。The traditional fire hazard identification method has low identification accuracy, large difference between the maximum heat release rate and the actual value, and high hazard growth coefficient. This paper presents a method of laboratory fire hazard identification based on machine vision. Firstly, the laboratory data were collected in detail. According to the results of outlier filtering and normal vector of the collected data, the data were smoothly de-noised. Secondly, the differential evolution method was introduced to optimize the online sequence limit-learning machine in machine vision. Then, the laboratory data after smooth de-noising were input into the optimized learning machine and trained. Finally, the laboratory fire hazard source was identified. The experimental results show that this method has high identification accuracy and low hazard growth coefficient, and the maximum heat release rate is close to the actual value.

关 键 词：机器视觉 极限学习机 实验室 火灾危险源 辨识方法 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]