摘要
针对无线多媒体传感网络在森林火灾监测应用中存在的问题,提出了基于图像哈希编码技术的森林火灾识别算法。首先,建立森林火灾图像的图像库,提取火焰图像的一系列静态和动态特征,通过哈希函数对其特征向量进行计算得到对应的哈希码,从而得到图像库对应的哈希码库。其次,计算被识别图像的哈希码,并通过计算汉明距离与哈希码库进行匹配,得出与其最相近的图像,从而得出是否有火灾发生。实验结果表明,该算法的火焰识别准确率达到94.12%,高于SVM、BP神经网络和稀疏表示的火焰识别算法,且减少了网络中因图像传输而产生的能量消耗,提高了网络带宽的使用率。
Aiming at the existing problems about the application of the wireless multimedia sensor network(WMSN)in the forest fire monitoring,we proposed a forest fire recognition algorithm based on the mage hash code technology.First,we built the database of forest fire image and got database of corresponding hash code.Through extracting a series of static and dynamic characteristics of the flame image,we computed characteristic vector generation with the hash function to get the corresponding hash code.Second,we computed the hash code of the identified image,matched the image by calculating the hamming distance between the identified image and the image of database,and got the most similar image with the identified image,obtaining the conclusion of the presence of fire or non-fire.The experimental results show that the accuracy of flame recognition of this algorithm is 94.12%,and it is higher than other flame recognition algorithms based on the SVM,BP-neural network and sparse representation.Besides,this algorithm reduces the energy consumption of image transmission in WMSN,and improves the use efficiency of the network bandwidth.
出处
《计算机科学》
CSCD
北大核心
2016年第5期313-317,共5页
Computer Science
基金
国家自然科学基金(61202163
61373100)
山西省科技攻关项目(20120313032-3)
虚拟现实技术与系统国家重点实验室(BUAA-VR-15KF02)资助
关键词
林火监测
无线多媒体传感网络
火焰识别
图像哈希编码
Forest fire monitoring
Wireless media sensor network
Fire recognition
Image hash coding