on the basis of analyzing the characteristics of low light level(LLL)image and ultra-violet image and the information amount of dual channel color night vision system,the LLL and ultra-violet color night vision techni...on the basis of analyzing the characteristics of low light level(LLL)image and ultra-violet image and the information amount of dual channel color night vision system,the LLL and ultra-violet color night vision technique is put forward.The methods of gray-scale modulation,frequency field fusion,special component fusion arc tried,and the improved LLL and ultra-violet image pseudo color fusion algorithms are presented.These new algorithms include subsection gray-scale modulation,image difference picking-up,component separation based on the reflected characteristics to night skylight reflection characteristics of objects and color space mapping which embodies the spectrum response of image sensor and nature vision.Some good results are obtained.展开更多
微光/红外图像彩色融合是目前国内外夜视技术的重要发展方向,在超低照度下(环境照度小于2×10-3 lux),由于成像器件限制,微光图像具有低信噪比、低对比度等特点,导致目标难以辨识,成为制约彩色夜视技术的关键。为了提高目标的探测...微光/红外图像彩色融合是目前国内外夜视技术的重要发展方向,在超低照度下(环境照度小于2×10-3 lux),由于成像器件限制,微光图像具有低信噪比、低对比度等特点,导致目标难以辨识,成为制约彩色夜视技术的关键。为了提高目标的探测和识别率,提出了一种基于卷积自编码网络的微光图像复原方法,利用卷积自编码网络从微光图像训练集中学习超低照度下微光图像特征,实现去噪和对比度增强。实验结果表明,本文提出的方法得到的峰值信噪比(Peak Signal to Noise Ratio,PSNR)较经典的BM3D算法平均提高1.67dB,结构相似度(Structural Similarity Index,SSIM)的值平均提高0.063,均方根对比度的值(Root Mean Square Contrast,RMSC)平均提高0.19。对微光图像复原具有很好的效果,能够有效地提高信噪比和对比度水平。展开更多
文摘on the basis of analyzing the characteristics of low light level(LLL)image and ultra-violet image and the information amount of dual channel color night vision system,the LLL and ultra-violet color night vision technique is put forward.The methods of gray-scale modulation,frequency field fusion,special component fusion arc tried,and the improved LLL and ultra-violet image pseudo color fusion algorithms are presented.These new algorithms include subsection gray-scale modulation,image difference picking-up,component separation based on the reflected characteristics to night skylight reflection characteristics of objects and color space mapping which embodies the spectrum response of image sensor and nature vision.Some good results are obtained.
文摘微光/红外图像彩色融合是目前国内外夜视技术的重要发展方向,在超低照度下(环境照度小于2×10-3 lux),由于成像器件限制,微光图像具有低信噪比、低对比度等特点,导致目标难以辨识,成为制约彩色夜视技术的关键。为了提高目标的探测和识别率,提出了一种基于卷积自编码网络的微光图像复原方法,利用卷积自编码网络从微光图像训练集中学习超低照度下微光图像特征,实现去噪和对比度增强。实验结果表明,本文提出的方法得到的峰值信噪比(Peak Signal to Noise Ratio,PSNR)较经典的BM3D算法平均提高1.67dB,结构相似度(Structural Similarity Index,SSIM)的值平均提高0.063,均方根对比度的值(Root Mean Square Contrast,RMSC)平均提高0.19。对微光图像复原具有很好的效果,能够有效地提高信噪比和对比度水平。
