Although compressive measurements save data storage and bandwidth usage, they are difficult to be used directly for target tracking and classification without pixel reconstruction. This is because the Gaussian random ...Although compressive measurements save data storage and bandwidth usage, they are difficult to be used directly for target tracking and classification without pixel reconstruction. This is because the Gaussian random matrix destroys the target location information in the original video frames. This paper summarizes our research effort on target tracking and classification directly in the compressive measurement domain. We focus on one particular type of compressive measurement using pixel subsampling. That is, original pixels in video frames are randomly subsampled. Even in such a special compressive sensing setting, conventional trackers do not work in a satisfactory manner. We propose a deep learning approach that integrates YOLO (You Only Look Once) and ResNet (residual network) for multiple target tracking and classification. YOLO is used for multiple target tracking and ResNet is for target classification. Extensive experiments using short wave infrared (SWIR), mid-wave infrared (MWIR), and long-wave infrared (LWIR) videos demonstrated the efficacy of the proposed approach even though the training data are very scarce.展开更多
Pixel-wise Code Exposure (PCE) camera is one type of compressive sensing camera that has low power consumption and high compression ratio. Moreover, a PCE camera can control individual pixel exposure time that can ena...Pixel-wise Code Exposure (PCE) camera is one type of compressive sensing camera that has low power consumption and high compression ratio. Moreover, a PCE camera can control individual pixel exposure time that can enable high dynamic range. Conventional approaches of using PCE camera involve a time consuming and lossy process to reconstruct the original frames and then use those frames for target tracking and classification. In this paper, we present a deep learning approach that directly performs target tracking and classification in the compressive measurement domain without any frame reconstruction. Our approach has two parts: tracking and classification. The tracking has been done using YOLO (You Only Look Once) and the classification is achieved using Residual Network (ResNet). Extensive experiments using mid-wave infrared (MWIR) and long-wave infrared (LWIR) videos demonstrated the efficacy of our proposed approach.展开更多
We report a new nBn photodetector(nBn-PD)design based on the InAlSb/AlSb/InAlSb/InAsSb material systems for midwavelength infrared(MWIR)applications.In this structure,delta-doped compositionally graded barrier(δ-DCGB...We report a new nBn photodetector(nBn-PD)design based on the InAlSb/AlSb/InAlSb/InAsSb material systems for midwavelength infrared(MWIR)applications.In this structure,delta-doped compositionally graded barrier(δ-DCGB)layers are suggested,the advantage of which is creation of a near zero valence band ofset in nBn photodetectors.The design of theδ-DCGB nBn-PD device includes a 3µm absorber layer(n-InAs0.81Sb0.19),a unipolar barrier layer(AlSb),and 0.2μm contact layer(n-InAs0.81Sb0.19)as well as a 0.116µm linear grading region(InAlSb)from the contact to the barrier layer and also from the barrier to the absorber layer.The analysis includes various dark current contributions,such as the Shockley-Read-Hall(SRH),trap-assisted tunneling(TAT),Auger,and Radiative recombination mechanisms,to acquire more precise results.Consequently,we show that the method used in the nBn device design leads to difusion-limited dark current so that the dark current density is 2.596×10^(−8)A/cm^(2)at 150 K and a bias voltage of−0.2 V.The proposed nBn detector exhibits a 50%cutof wavelength of more than 5µm,the peak current responsivity is 1.6 A/W at a wavelength of 4.5µm and a−0.2 V bias with 0.05 W/cm2 backside illumination without anti-refective coating.The maximum quantum efciency at 4.5µm is about 48.6%,and peak specifc detectivity(D*)is of 3.37×10^(10)cm⋅Hz1/2/W.Next,to solve the refection concern in this nBn devices,we use a BaF_(2)anti-refection coating layer due to its high transmittance in the MWIR window.It leads to an increase of almost 100%in the optical response metrics,such as the current responsivity,quantum efciency,and detectivity,compared to the optical response without an anti-refection coating layer.展开更多
文摘Although compressive measurements save data storage and bandwidth usage, they are difficult to be used directly for target tracking and classification without pixel reconstruction. This is because the Gaussian random matrix destroys the target location information in the original video frames. This paper summarizes our research effort on target tracking and classification directly in the compressive measurement domain. We focus on one particular type of compressive measurement using pixel subsampling. That is, original pixels in video frames are randomly subsampled. Even in such a special compressive sensing setting, conventional trackers do not work in a satisfactory manner. We propose a deep learning approach that integrates YOLO (You Only Look Once) and ResNet (residual network) for multiple target tracking and classification. YOLO is used for multiple target tracking and ResNet is for target classification. Extensive experiments using short wave infrared (SWIR), mid-wave infrared (MWIR), and long-wave infrared (LWIR) videos demonstrated the efficacy of the proposed approach even though the training data are very scarce.
文摘Pixel-wise Code Exposure (PCE) camera is one type of compressive sensing camera that has low power consumption and high compression ratio. Moreover, a PCE camera can control individual pixel exposure time that can enable high dynamic range. Conventional approaches of using PCE camera involve a time consuming and lossy process to reconstruct the original frames and then use those frames for target tracking and classification. In this paper, we present a deep learning approach that directly performs target tracking and classification in the compressive measurement domain without any frame reconstruction. Our approach has two parts: tracking and classification. The tracking has been done using YOLO (You Only Look Once) and the classification is achieved using Residual Network (ResNet). Extensive experiments using mid-wave infrared (MWIR) and long-wave infrared (LWIR) videos demonstrated the efficacy of our proposed approach.
文摘We report a new nBn photodetector(nBn-PD)design based on the InAlSb/AlSb/InAlSb/InAsSb material systems for midwavelength infrared(MWIR)applications.In this structure,delta-doped compositionally graded barrier(δ-DCGB)layers are suggested,the advantage of which is creation of a near zero valence band ofset in nBn photodetectors.The design of theδ-DCGB nBn-PD device includes a 3µm absorber layer(n-InAs0.81Sb0.19),a unipolar barrier layer(AlSb),and 0.2μm contact layer(n-InAs0.81Sb0.19)as well as a 0.116µm linear grading region(InAlSb)from the contact to the barrier layer and also from the barrier to the absorber layer.The analysis includes various dark current contributions,such as the Shockley-Read-Hall(SRH),trap-assisted tunneling(TAT),Auger,and Radiative recombination mechanisms,to acquire more precise results.Consequently,we show that the method used in the nBn device design leads to difusion-limited dark current so that the dark current density is 2.596×10^(−8)A/cm^(2)at 150 K and a bias voltage of−0.2 V.The proposed nBn detector exhibits a 50%cutof wavelength of more than 5µm,the peak current responsivity is 1.6 A/W at a wavelength of 4.5µm and a−0.2 V bias with 0.05 W/cm2 backside illumination without anti-refective coating.The maximum quantum efciency at 4.5µm is about 48.6%,and peak specifc detectivity(D*)is of 3.37×10^(10)cm⋅Hz1/2/W.Next,to solve the refection concern in this nBn devices,we use a BaF_(2)anti-refection coating layer due to its high transmittance in the MWIR window.It leads to an increase of almost 100%in the optical response metrics,such as the current responsivity,quantum efciency,and detectivity,compared to the optical response without an anti-refection coating layer.
