摘要
为了提高森林灭火航弹投放的准确性,提高航弹的利用率,对森林火灾图像识别方法进行了优化设计,将阈值分割和神经网络算法相结合,实现了火灾图像的智能识别,提高了图像识别的精确性。本研究利用高清摄像机获取森林火灾图像,采用阈值分割技术对图像进行提取,通过MATLAB计算得到有无火灾的信息,控制系统获取火灾信息后对灭火航弹的投放进行决策,提高了灭火过程的自动化水平。对图像识别算法进行了测试,通过测试发现,该算法可以有效地提高火灾的识别效率和识别精度,为森林灭火的自动化提供了参考。
In order to improve the release accuracy and utilization efficiency of fire extinguishing bomb, a forest fireimage recognition technique was designed. Threshold segmentation was combined with the neural network algorithmto realize the intelligent image recognition of forest fire, which improves the accuracy of image recognition. The im-age of forest fire was acquired from HD video cameras and extracted by the technology of threshold segmentation inthis study. The fire information can be obtained by MATLAB, and decisions on the release of fire extinguishingbombs were made by the control system after the fire information was analyzed. Therefore, the automatic level offirefighting was greatly improved. In addition, tests were made on the image recognition algorithm, and the resultsprove that the recognition efficiency and accuracy of forest fires can be effectively improved by using the proposedalgorithm, which provides a reference for the automatic forest firefighting.
出处
《应用科技》
CAS
2016年第3期82-86,共5页
Applied Science and Technology
基金
哈尔滨市应用技术研究与开发项目(2013RFXXJ002)
高等学校博士学科点专项科研基金项目(20102304110007)
关键词
森林灭火
神经网络
灭火航弹
阈值分割
图像识别
forest firefighting
neural network
fire extinguishing bomb
threshold segmentation
image recognition