Real-time detection of driver fatigue status is of great significance for road traffic safety.In this paper,a proposed novel driver fatigue detection method is able to detect the driver’s fatigue status around the cl...Real-time detection of driver fatigue status is of great significance for road traffic safety.In this paper,a proposed novel driver fatigue detection method is able to detect the driver’s fatigue status around the clock.The driver’s face images were captured by a camera with a colored lens and an infrared lens mounted above the dashboard.The landmarks of the driver’s face were labeled and the eye-area was segmented.By calculating the aspect ratios of the eyes,the duration of eye closure,frequency of blinks and PERCLOS of both colored and infrared,fatigue can be detected.Based on the change of light intensity detected by a photosensitive device,the weight matrix of the colored features and the infrared features was adjusted adaptively to reduce the impact of lighting on fatigue detection.Video samples of the driver’s face were recorded in the test vehicle.After training the classification model,the results showed that our method has high accuracy on driver fatigue detection in both daytime and nighttime.展开更多
A long-/long-wave dual-color detector with N-M-π-B-π-M-N structure was developed based on a type-Ⅱ InAs/GaSb superlattice. The saturated responsivity was achieved under low bias voltage for both channels. The devic...A long-/long-wave dual-color detector with N-M-π-B-π-M-N structure was developed based on a type-Ⅱ InAs/GaSb superlattice. The saturated responsivity was achieved under low bias voltage for both channels. The device could be operated as a single detector for sequential detection and showed high quantum efficiencies. The peak quantum efficiencies of long-wavelength infrared band-1(blue channel) and long-wavelength infrared band-2(red channel) were 44% at 6.3 μm under 20 mV and 57% at 9.1 μm under-60 mV, respectively. The optical performance for each channel was achieved using a 2 μm thickness absorber. Due to the high QE, the specific detectivities of the blue and red channels reached5.0×10^(11) cm·Hz^(1/2)/W at 6.8 μm and 3.1×10^(11) cm·Hz1^(1/2)/W at 9.1 μm, respectively, at 77 K.展开更多
Adding colors to monochrome thermal infrared images can help observers understand the scenery better. A nonlinear color estimation method for single-band thermal infrared imagery based on kernel principal component an...Adding colors to monochrome thermal infrared images can help observers understand the scenery better. A nonlinear color estimation method for single-band thermal infrared imagery based on kernel principal component analysis (KPCA) and sparse representation was proposed. Nonlinear features of infrared image were extracted using KPCA. The relationship between image features and chromatic values was learned using sparse representation and a color estimation model was obtained. The thermal infrared images can be rendered automatically using the color estimation model. The experimental results show that the proposed method can render infrared image with an accurate color appearance. The proposed idea can also be used in other color estimation problem.展开更多
The use of Infrared Thermal Scanners proved to be very useful in lots of applications. Using different color palettes, temperatures can be well-represented in the resulting image. However, most color palettes in hot t...The use of Infrared Thermal Scanners proved to be very useful in lots of applications. Using different color palettes, temperatures can be well-represented in the resulting image. However, most color palettes in hot tropical places like the Philippines are unsuitable since the ambient temperature is almost the same as the scanned object or person. This study evaluates twelve (12) known and used color palettes in the market to determine the most suitable for tropical places using the edge/border tracing algorithms Sobel-Feldman and Laplacian. The result shows that color palettes with the most colors produce more noise, making it difficult to distinguish the object scanned from the background. On the other hand, color palettes with three or fewer contrasting colors produce crisp and more detailed results. This study helps developers and researchers efficiently use color combinations suitable for hot weather for an effective thermal scanning and image representation.展开更多
目的利用近红外光谱及光谱融合策略,结合化学计量学方法,建立水性油墨的颜色预测模型,实现水性油墨印刷品颜色准确预测。方法采集不同酒精含量和不同调色墨含量的油墨的近红外光谱反射率和吸光度数据,并测得对应的印刷品的Lab值,然后建...目的利用近红外光谱及光谱融合策略,结合化学计量学方法,建立水性油墨的颜色预测模型,实现水性油墨印刷品颜色准确预测。方法采集不同酒精含量和不同调色墨含量的油墨的近红外光谱反射率和吸光度数据,并测得对应的印刷品的Lab值,然后建立单一光谱不同预处理过后的偏最小二乘(Partial Least Squares,PLS)模型,以及基于数据层融合和特征层融合的PLS模型,最终通过比较预测集决定系数和预测集均方根误差(Root Mean Square Error of Prediction,RMSEP)以及色差来评估模型的预测效果。结果单光谱建模,基于反射率建立的模型准确率高于基于吸光度建立的模型;数据层融合缺乏稳定性,对L和b值的预测有所提升,对a值的预测几乎不变;特征层融合建模效果明显好于单一光谱和数据层融合,对Lab的预测决定系数分别达到了0.9961、0.9939、0.9974;RMSEP值分别为0.1421、0.2126、0.2072;预测值与真实值的最大色差为0.6783。结论通过光谱特征融合技术能提高油墨颜色预测精度,准确预测出油墨颜色变化。展开更多
为了解决红外制导研究中舰船图像样本数量不足的问题,提出一种面向舰船图像的改进的生成对抗网络(generative adversarial network,GAN),能够生成高质量的红外图像。首先转换可见光图像颜色空间以更好地捕捉夜间低亮度下图像的轮廓信息...为了解决红外制导研究中舰船图像样本数量不足的问题,提出一种面向舰船图像的改进的生成对抗网络(generative adversarial network,GAN),能够生成高质量的红外图像。首先转换可见光图像颜色空间以更好地捕捉夜间低亮度下图像的轮廓信息,然后引入残差块生成网络降低低像素的可见光图像对生成的红外图像的影响并加深网络层数以更好地学习深层映射关系,最后引入更平滑的损失函数加快收敛速度,提高生成红外图像目标边缘清晰程度。在制作的无人机拍摄的红外可见光配对的数据集进行测试,改进后的方法平均生成图像峰值信噪比(peak signal to noiseratio,PSNR)提升20.3%,结构相似性度量(structural similarity,SSIM)提升30.4%。结果表明改进的网络可以生成质量更高的红外仿真图像,用于目标检测等任务有更好的效果。展开更多
基金The work of this paper was supported by the National Natural Science Foundation of China under grant numbers 61572038 received by J.Z.in 2015.URL:https://isisn.nsfc.gov.cn/egrantindex/funcindex/prjsearch-list。
文摘Real-time detection of driver fatigue status is of great significance for road traffic safety.In this paper,a proposed novel driver fatigue detection method is able to detect the driver’s fatigue status around the clock.The driver’s face images were captured by a camera with a colored lens and an infrared lens mounted above the dashboard.The landmarks of the driver’s face were labeled and the eye-area was segmented.By calculating the aspect ratios of the eyes,the duration of eye closure,frequency of blinks and PERCLOS of both colored and infrared,fatigue can be detected.Based on the change of light intensity detected by a photosensitive device,the weight matrix of the colored features and the infrared features was adjusted adaptively to reduce the impact of lighting on fatigue detection.Video samples of the driver’s face were recorded in the test vehicle.After training the classification model,the results showed that our method has high accuracy on driver fatigue detection in both daytime and nighttime.
基金supported by the National Key Technology R&D Program of China(Grant Nos.2018YFA0209104 and 2016YFB0402403)
文摘A long-/long-wave dual-color detector with N-M-π-B-π-M-N structure was developed based on a type-Ⅱ InAs/GaSb superlattice. The saturated responsivity was achieved under low bias voltage for both channels. The device could be operated as a single detector for sequential detection and showed high quantum efficiencies. The peak quantum efficiencies of long-wavelength infrared band-1(blue channel) and long-wavelength infrared band-2(red channel) were 44% at 6.3 μm under 20 mV and 57% at 9.1 μm under-60 mV, respectively. The optical performance for each channel was achieved using a 2 μm thickness absorber. Due to the high QE, the specific detectivities of the blue and red channels reached5.0×10^(11) cm·Hz^(1/2)/W at 6.8 μm and 3.1×10^(11) cm·Hz1^(1/2)/W at 9.1 μm, respectively, at 77 K.
基金National Natural Science Foundation of China(No. 61072090)the Fundamental Research Funds for the Central Universities,China+2 种基金Shanghai Pujiang Program,China(No. 12PJ1402200)China Postdoctoral Science Foundation Funded Project(No. 2012M511058)Shanghai Postdoctoral Sustentation Fund,China(No. 12R21412500)
文摘Adding colors to monochrome thermal infrared images can help observers understand the scenery better. A nonlinear color estimation method for single-band thermal infrared imagery based on kernel principal component analysis (KPCA) and sparse representation was proposed. Nonlinear features of infrared image were extracted using KPCA. The relationship between image features and chromatic values was learned using sparse representation and a color estimation model was obtained. The thermal infrared images can be rendered automatically using the color estimation model. The experimental results show that the proposed method can render infrared image with an accurate color appearance. The proposed idea can also be used in other color estimation problem.
文摘The use of Infrared Thermal Scanners proved to be very useful in lots of applications. Using different color palettes, temperatures can be well-represented in the resulting image. However, most color palettes in hot tropical places like the Philippines are unsuitable since the ambient temperature is almost the same as the scanned object or person. This study evaluates twelve (12) known and used color palettes in the market to determine the most suitable for tropical places using the edge/border tracing algorithms Sobel-Feldman and Laplacian. The result shows that color palettes with the most colors produce more noise, making it difficult to distinguish the object scanned from the background. On the other hand, color palettes with three or fewer contrasting colors produce crisp and more detailed results. This study helps developers and researchers efficiently use color combinations suitable for hot weather for an effective thermal scanning and image representation.
文摘目的利用近红外光谱及光谱融合策略,结合化学计量学方法,建立水性油墨的颜色预测模型,实现水性油墨印刷品颜色准确预测。方法采集不同酒精含量和不同调色墨含量的油墨的近红外光谱反射率和吸光度数据,并测得对应的印刷品的Lab值,然后建立单一光谱不同预处理过后的偏最小二乘(Partial Least Squares,PLS)模型,以及基于数据层融合和特征层融合的PLS模型,最终通过比较预测集决定系数和预测集均方根误差(Root Mean Square Error of Prediction,RMSEP)以及色差来评估模型的预测效果。结果单光谱建模,基于反射率建立的模型准确率高于基于吸光度建立的模型;数据层融合缺乏稳定性,对L和b值的预测有所提升,对a值的预测几乎不变;特征层融合建模效果明显好于单一光谱和数据层融合,对Lab的预测决定系数分别达到了0.9961、0.9939、0.9974;RMSEP值分别为0.1421、0.2126、0.2072;预测值与真实值的最大色差为0.6783。结论通过光谱特征融合技术能提高油墨颜色预测精度,准确预测出油墨颜色变化。
文摘为了解决红外制导研究中舰船图像样本数量不足的问题,提出一种面向舰船图像的改进的生成对抗网络(generative adversarial network,GAN),能够生成高质量的红外图像。首先转换可见光图像颜色空间以更好地捕捉夜间低亮度下图像的轮廓信息,然后引入残差块生成网络降低低像素的可见光图像对生成的红外图像的影响并加深网络层数以更好地学习深层映射关系,最后引入更平滑的损失函数加快收敛速度,提高生成红外图像目标边缘清晰程度。在制作的无人机拍摄的红外可见光配对的数据集进行测试,改进后的方法平均生成图像峰值信噪比(peak signal to noiseratio,PSNR)提升20.3%,结构相似性度量(structural similarity,SSIM)提升30.4%。结果表明改进的网络可以生成质量更高的红外仿真图像,用于目标检测等任务有更好的效果。