基于小波分析的快照式高光谱海面溢油厚度分析

Snapshot Hyperspectral Analysis of Spilled Oil Thickness Based on Wavelet Analysis

针对海面溢油厚度的探测需求,通过模拟海面环境,采用快照式高光谱相机对不同厚度油膜的反射光谱进行探测;首先根据快照式高光谱相机获得的高光谱图像,计算得到了不同厚度油膜的反射光谱;然后采用小波变换对反射光谱进行分析,再利用Coif5小波基进行离散小波变换得到第9层重构信号的细节系数,用其反映油膜厚度的差异。油膜厚度和细节系数具有良好的线性关系,据此可准确识别大范围溢油事故中不同位置油膜的厚度。快照式高光谱探测方式极大地提高了溢油分析的效率,对溢油灾害的快速发现和实时监测具有参考价值。

Herein,the sea surface environment was simulated and a snapshot hyperspectral camera was used to detect the reflection spectra of spilled oil films with different thicknesses at the sea surface.First,based on hyperspectral images obtained using the snapshot hyperspectral camera,the reflection spectra of oil films were obtained.Hereafter,the wavelet transform was used for analysis,and the Coif5 wavelet base was used as the discrete wavelet transform to obtain the detail coefficient of the 9 th layer reconstructed signal to determine the difference in oil film thickness.The oil film thickness and detail coefficient exhibited a good linear relationship,thus enabling accurate identification of the oil film thickness at different locations in a large-scale spilled oil accident.The snapshot hyperspectral detection method used herein significantly improves the spilled oil analysis efficiency and provides references for the rapid detection and real-time monitoring of spilled oil disasters.