摘要
由于利用边缘特征点进行目标识别时识别误差较大,现有数学建模方法存在直线拟合能力较差的问题,提出一种基于压缩感知的电气设备红外图像目标检测数学建模方法。对电气设备红外图像进行预处理,处理步骤包括图像灰度化与图像平滑。通过压缩感知获取电气设备红外图像目标的可压缩与稀疏信号,利用数学形态运算对目标信号的对应形状进行提取与度量,实现图像目标的识别。为了证明提出的数学建模方法的直线拟合能力较好,将现有的3种数学建模方法作为对比实验方法,进行直线拟合能力的对比实验,实验结果证明提出的数学建模方法在电气设备a相、b相、c相的直线拟合率均高于其他实验方法,更适合应用于电气设备红外图像目标检测的数学建模。
Due to the large recognition error when using edge feature points for target recognition,the existing mathematical modeling methods have the problem of poor line fitting ability.A mathematical modeling method based on compressed sensing for infrared image target detection of electrical equipment is proposed.The infrared image of electrical equipment is preprocessed,and the processing steps include image graying and image smoothing.The compressed and sparse signals of the infrared image target of the electrical equipment are obtained by the compressed sensing,and the corresponding shape of the target signal is extracted and measured by the mathematical morphological operation to realize the recognition of the image target.In order to prove the better linear fitting ability of the proposed mathematical modeling method,three existing mathematical modeling methods are used as comparative experimental methods to carry out the comparative experiment of the linear fitting ability.The experimental results show that the linear fitting rate of the proposed mathematical modeling method in phase a,phase B and phase C of electrical equipment is higher than that of other experimental methods,which is more suitable for mathematical modeling of the electrical equipment infrared image target detection.
作者
廖红菊
LIAO Hongju(Enshi Vocational and Technical College,Enshi Hubei 445000,China)
出处
《自动化与仪器仪表》
2020年第10期47-50,共4页
Automation & Instrumentation
基金
湖北省职业技术教育学会课题:小学数学教学设计与实施课程“46634”教学模式的研究与实践(No.ZJGA201843)。
关键词
压缩感知
电气设备
红外图像
目标检测
数学建模
compressed sensing
electrical equipment
infrared image
target detection
mathematical modeling
