Probability-Enhanced Anchor-Free Detector for Remote-Sensing Object Detection

Anchor-free object-detection methods achieve a significant advancement in field of computer vision,particularly in the realm of real-time inferences.However,in remote sensing object detection,anchor-free methods often lack of capability in separating the foreground and background.This paper proposes an anchor-free method named probability-enhanced anchor-free detector(ProEnDet)for remote sensing object detection.First,a weighted bidirectional feature pyramid is used for feature extraction.Second,we introduce probability enhancement to strengthen the classification of the object’s foreground and background.The detector uses the logarithm likelihood as the final score to improve the classification of the foreground and background of the object.ProEnDet is verified using the DIOR and NWPU-VHR-10 datasets.The experiment achieved mean average precisions of 61.4 and 69.0 on the DIOR dataset and NWPU-VHR-10 dataset,respectively.ProEnDet achieves a speed of 32.4 FPS on the DIOR dataset,which satisfies the real-time requirements for remote-sensing object detection.

基于anchor-free的光学遥感舰船关重部位检测算法

针对基于深度学习的遥感舰船检测算法存在精细化程度不足、检测效率低的问题,提出一种基于anchor-free的光学遥感舰船关重部位检测算法。所提算法以全卷积的单阶段目标检测(FCOS)算法为基准,在主干网络中引入全局上下文模块,提高网络的特征表达能力;为更好地描述目标的方向性,在预测阶段构建了具有方向表征能力的回归分支;对中心度函数进行优化,使其具备方向感知和自适应能力。实验结果表明:在自建舰船关重部位数据集和HRSC2016上,所提算法的平均精度(AP)比FCOS算法有显著提升;与其他算法相比,所提算法在检测速度和检测精度上均表现优越,具有较高的检测效率。

Current-Aided Multiple-AUV Cooperative Localization and Target Tracking in Anchor-Free Environments

In anchor-free environments,where no devices with known positions are available,the error growth of autonomous underwater vehicle(AUV)localization and target tracking is unbounded due to the lack of references and the accumulated errors in inertial measurements.This paper aims to improve the localization and tracking accuracy by involving current information as extra references.We first integrate current measurements and maps with belief propagation and design a distributed current-aided message-passing scheme that theoretically solves the localization and tracking problems.Based on this scheme,we propose particle-based cooperative localization and target tracking algorithms,named CaCL and CaTT,respectively.In AUV localization,CaCL uses the current measurements to correct the predicted and transmitted position information and alleviates the impact of the accumulated errors in inertial measurements.With target tracking,the current maps are applied in CaTT to modify the position prediction of the target which is calculated through historical estimates.The effectiveness and robustness of the proposed methods are validated through various simulations by comparisons with alternative methods under different trajectories and current conditions.

基于Anchor-free的孪生网络目标跟踪算法

基于anchor-free的目标预测方法相较于anchor-based方法速度更快,故设计一种基于anchor-free的孪生网络目标跟踪算法(AFSN)。通过对特征图、预测结果进行双重融合来提升跟踪效果,由深至浅对特征图进行堆叠融合,利用多层特征图进行目标预测,融合多个预测结果来稳定跟踪效果。采用anchor-free的目标预测方法,直接在像素点上进行目标类别的预测和边界框回归,避免了需设计大量锚点包围盒的问题。在GOT-10K数据集上,该算法的平均重叠率(AO)和成功率(SR_(0.75))相较于SiamRPN++算法提高了4.9和9.9百分点,算法处理速度可达每秒37帧。

PTW Detector7293种剂量验证方法在调强放射治疗中的对比研究

目的 :对比PTW Detector729 3种验证方法在调强放射治疗(intensity modulated radiation therapy,IMRT)剂量验证中的差异。方法:回顾性选取2022年1—12月于某院完成放射治疗的鼻咽癌、肺癌、乳腺癌、宫颈癌和全脑放射治疗患者共50例,使用PTW Detector729二维电离室矩阵配合PTW RW3固体水与PTW Ocavius 4D旋转模体对50例患者的IMRT计划分别进行归零机架角2D剂量验证、实际机架角2D剂量验证和实际机架角3D剂量验证。将剂量评估阈值设为10%,统计3种验证方法在3%/1 mm、2%/2 mm、3%/2 mm和3%/3 mm 4种评估标准下的γ通过率。采用SPSS 22.0统计学软件进行数据分析。结果:在10%剂量评估阈值标准下,归零机架角2D剂量验证的γ通过率最高,差异有统计学意义(P<0.05);实际机架角2D剂量验证的γ通过率高于实际机架角3D剂量验证,差异有统计学意义(P<0.05)。3种验证方法的γ通过率在3%/1 mm、2%/2 mm、3%/2 mm和3%/3 mm 4种标准下逐渐升高,且在3%/2 mm标准下均超过90%,结果满足临床放射治疗要求。结论:3种验证方法的验证结果均能达到IMRT剂量验证实践指南要求,根据IMRT开展情况选择合适的验证方法,对确保治疗方案安全、有效实施具有重要意义。

DAAPS: A Deformable-Attention-Based Anchor-Free Person Search Model

Person Search is a task involving pedestrian detection and person re-identification,aiming to retrieve person images matching a given objective attribute from a large-scale image library.The Person Search models need to understand and capture the detailed features and context information of smaller objects in the image more accurately and comprehensively.The current popular Person Search models,whether end-to-end or two-step,are based on anchor boxes.However,due to the limitations of the anchor itself,the model inevitably has some disadvantages,such as unbalance of positive and negative samples and redundant calculation,which will affect the performance of models.To address the problem of fine-grained understanding of target pedestrians in complex scenes and small sizes,this paper proposes a Deformable-Attention-based Anchor-free Person Search model(DAAPS).Fully Convolutional One-Stage(FCOS),as a classic Anchor-free detector,is chosen as the model’s infrastructure.The DAAPS model is the first to combine the Anchor-free Person Search model with Deformable Attention Mechanism,applied to guide the model adaptively adjust the perceptual.The Deformable Attention Mechanism is used to help the model focus on the critical information and effectively improve the poor accuracy caused by the absence of anchor boxes.The experiment proves the adaptability of the Attention mechanism to the Anchor-free model.Besides,with an improved ResNeXt+network frame,the DAAPS model selects the Triplet-based Online Instance Matching(TOIM)Loss function to achieve a more precise end-to-end Person Search task.Simulation experiments demonstrate that the proposed model has higher accuracy and better robustness than most Person Search models,reaching 95.0%of mean Average Precision(mAP)and 95.6%of Top-1 on the CUHK-SYSU dataset,48.6%of mAP and 84.7%of Top-1 on the Person Re-identification in the Wild(PRW)dataset,respectively.

Low-noise and low-power pixel sensor chip for gas pixel detectors

Topmetal-M2 is a large-area pixel sensor chip fabricated using the GSMC 130 nm CMOS process in 2021.The pixel array of Topmetal-M2 consists of pixels of 400 rows×512 columns with a pixel pitch of 45μm×45μm.The array is divided into 16 subarrays,with pixels of 400 rows×32 columns per subarray.Each pixel incorporates two charge sensors:a diode sensor and a Topmetal sensor.The in-pixel circuit primarily consists of a charge-sensitive amplifier for energy measurements,a discriminator with a peak-holding circuit,and a time-to-amplitude converter for time-of-arrival measurements.The pixel of Topmetal-M2 has a charge input range of~0-3 k e-,a voltage output range of~0-180 mV,and a charge-voltage conversion gain of~59.56μV∕e-.The average equivalent noise charge of Topmetal-M2,which includes the readout electronic system noise,is~43.45 e-.In the scanning mode,the time resolution of Topmetal-M2 is 1 LSB=1.25μs,and the precision is^()7.41μs.At an operating voltage of 1.5 V,Topmetal-M2 has a power consumption of~49 mW∕cm~2.In this article,we provide a comprehensive overview of the chip architecture,pixel working principles,and functional behavior of Topmetal-M2.Furthermore,we present the results of preliminary tests conducted on Topmetal-M2,namely,alpha-particle and soft X-ray tests.

CVTD: A Robust Car-Mounted Video Text Detector

Text perception is crucial for understanding the semantics of outdoor scenes,making it a key requirement for building intelligent systems for driver assistance or autonomous driving.Text information in car-mounted videos can assist drivers in making decisions.However,Car-mounted video text images pose challenges such as complex backgrounds,small fonts,and the need for real-time detection.We proposed a robust Car-mounted Video Text Detector(CVTD).It is a lightweight text detection model based on ResNet18 for feature extraction,capable of detecting text in arbitrary shapes.Our model efficiently extracted global text positions through the Coordinate Attention Threshold Activation(CATA)and enhanced the representation capability through stacking two Feature Pyramid Enhancement Fusion Modules(FPEFM),strengthening feature representation,and integrating text local features and global position information,reinforcing the representation capability of the CVTD model.The enhanced feature maps,when acted upon by Text Activation Maps(TAM),effectively distinguished text foreground from non-text regions.Additionally,we collected and annotated a dataset containing 2200 images of Car-mounted Video Text(CVT)under various road conditions for training and evaluating our model’s performance.We further tested our model on four other challenging public natural scene text detection benchmark datasets,demonstrating its strong generalization ability and real-time detection speed.This model holds potential for practical applications in real-world scenarios.

Flexibility potential of Cs_(2)BX_(6)(B=Hf,Sn,Pt,Zr,Ti;X=I,Br,Cl)with application in photovoltaic devices and radiation detectors

As interest in double perovskites is growing,especially in applications like photovoltaic devices,understanding their mechanical properties is vital for device durability.Despite extensive exploration of structure and optical properties,research on mechanical aspects is limited.This article builds a vacancyordered double perovskite model,employing first-principles calculations to analyze mechanical,bonding,electronic,and optical properties.Results show Cs_(2)Hfl_(6),Cs_(2)SnBr_(6),Cs_(2)SnI_(6),and Cs_(2)PtBr_(6)have Young's moduli below 13 GPa,indicating flexibility.Geometric parameters explain flexibility variations with the changes of B and X site composition.Bonding characteristic exploration reveals the influence of B and X site electronegativity on mechanical strength.Cs_(2)SnBr_(6)and Cs_(2)PtBr_(6)are suitable for solar cells,while Cs_(2)HfI_(6)and Cs_(2)TiCl_(6)show potential for semi-transparent solar cells.Optical property calculations highlight the high light absorption coefficients of up to 3.5×10^(5) cm^(-1)for Cs_(2)HfI_(6)and Cs_(2)TiCl_(6).Solar cell simulation shows Cs_(2)PtBr_(6)achieves 22.4%of conversion effciency.Cs_(2)ZrCl_(6)holds promise for ionizing radiation detection with its 3.68 eV bandgap and high absorption coefficient.Vacancy-ordered double perovskites offer superior flexibility,providing valuable insights for designing stable and flexible devices.This understanding enhances the development of functional devices based on these perovskites,especially for applications requiring high stability and flexibility.

Design and performance evaluation of a large field-of-view dual-particle time-encoded imager based on a depth-of-interaction detector

Time-encoded imaging is useful for identifying potential special nuclear materials and other radioactive sources at a distance.In this study,a large field-of-view time-encoded imager was developed for gamma-ray and neutron source hotspot imaging based on a depth-of-interaction(DOI)detector.The imager primarily consists of a DOI detector system and a rotary dual-layer cylindrical coded mask.An EJ276 plastic scintillator coupled with two SiPMs was designed as the DOI detector to increase the field of view and improve the imager performance.The difference in signal time at both ends and the log of the signal amplitude ratio were used to calculate the interaction position resolution.The position resolution of the DOI detector was calibrated using a collimated Cs-137 source,and the full width at half maximum of the reconstruction position of the Gaussian fitting curve was approximately 4.4 cm.The DOI detector can be arbitrarily divided into several units to independently reconstruct the source distribution images.The unit length was optimized via Am-Be source-location experiments.A multidetector filtering method is proposed for image denoising.This method can effectively reduce image noise caused by poor DOI detector position resolution.The vertical field of view of the imager was(-55°,55°)when the detector was placed in the center of the coded mask.A DT neutron source at 20 m standoff could be located within 2400 s with an angular resolution of 3.5°.

Self-powered UVC detectors based on α-Ga_(2)O_(3) with enchanted speed performance

Detectors were developed for detecting irradiation in the short-wavelength ultraviolet(UVC)interval using high-quality single-crystallineα-Ga_(2)O_(3) films with Pt interdigital contacts.The films ofα-Ga_(2)O_(3) were grown on planar sapphire substrates with c-plane orientation using halide vapor phase epitaxy.The spectral dependencies of the photo to dark current ratio,responsivity,external quantum efficiency and detectivity of the structures were investigated in the wavelength interval of 200−370 nm.The maximum of photo to dark current ratio,responsivity,external quantum efficiency,and detectivity of the structures were 1.16×10^(4) arb.un.,30.6 A/W,1.65×10^(4)%,and 6.95×10^(15) Hz^(0.5)·cm/W at a wavelength of 230 nm and an applied voltage of 1 V.The high values of photoelectric properties were due to the internal enhancement of the photoresponse associated with strong hole trapping.Theα-Ga_(2)O_(3) film-based UVC detectors can function in self-powered operation mode due to the built-in electric field at the Pt/α-Ga_(2)O_(3) interfaces.At a wavelength of 254 nm and zero applied voltage,the structures exhibit a responsivity of 0.13 mA/W and an external quantum efficiency of 6.2×10^(−2)%.The UVC detectors based on theα-Ga_(2)O_(3) films demonstrate high-speed performance with a rise time of 18 ms in self-powered mode.

Gas microchannel plate-pixel detector for X-ray polarimetry

POLAR-2 is a gamma-ray burst(GRB)polarimeter that is designed to study the polarization in GRB radiation emissions,aiming to improve our knowledge of related mechanisms.POLAR-2 is expected to utilize an on-board polarimeter that is sensitive to soft X-rays(2-10 keV),called low-energy polarization detector.We have developed a new soft X-ray polari-zation detector prototype based on gas microchannel plates(GMCPs)and pixel chips(Topmetal).The GMCPs have bulk resistance,which prevents charging-up effects and ensures gain stability during operation.The detector is composed of low outgassing materials and is gas-sealed using a laser welding technique,ensuring long-term stability.A modulation factor of 41.28%±0.64% is obtained for a 4.5 keV polarized X-ray beam.A residual modulation of 1.96%±0.58% at 5.9 keV is observed for the entire sensitive area.

Effect of multiple coulomb scattering on the beam tests of silicon pixel detectors

In the research and development of new silicon pixel detectors,a collimated monoenergetic charged-particle test beam equipped with a high-resolution pixel-beam telescope is crucial for prototype verification and performance evaluation.When the beam energy is low,the effect of multiple Coulomb scattering on the measured resolution of the Device Under Test(DUT)must be considered to accurately evaluate the performance of the pixel chips and detectors.This study aimed to investigate the effect of multiple Coulomb scattering on the measured resolution,particularly at low beam energies.Simulations were conducted using Allpix^(2) to study the effects of multiple Coulomb scattering under different beam energies,material budgets,and telescope layouts.The simulations also provided the minimum energy at which the effect of multiple Coulomb scattering could be ignored.Compared with the results of a five-layer detector system tested with an electron beam at DESY,the simulation results were consistent with the beam test results,confirming the reliability of the simulations.

Study on the Performance of the GRANDProto300 Particle Detector Array by Simulation

The Giant Radio Array for Neutrino Detection(GRAND)is a proposed large-scale observatory designed to detect cosmic rays,gamma-rays,and neutrinos with energies exceeding 100 Pe V.The GRANDProto300 experiment is proposed as the early stage of the GRAND project,consisting of a hybrid array of radio antennas and scintillator detectors.The latter,as a mature and traditional detector,is used to cross-check the nature of the candidate events selected from radio observations.In this study,we developed a simulation software called G4GRANDProto300,based on the Geant4 software package,to optimize the spacing of the scintillator detector array and to investigate its effective area.The analysis was conducted at various zenith angles under different detector spacings,including 300,500,600,700,and 900 m.Our results indicate that,for large zenith angles used to search for cosmic-ray in the GRAND project,the optimized effective area is with a detector spacing of 500 m.The G4GRANDProto300 software that we developed could be used to further optimize the layout of the particle detector array in future work.

Metal-Halide Perovskite Submicrometer-Thick Films for Ultra-Stable Self-Powered Direct X-Ray Detectors

Metal-halide perovskites are revolutionizing the world of X-ray detectors,due to the development of sensitive,fast,and cost-effective devices.Self-powered operation,ensuring portability and low power consumption,has also been recently demonstrated in both bulk materials and thin films.However,the signal stability and repeatability under continuous X-ray exposure has only been tested up to a few hours,often reporting degradation of the detection performance.Here it is shown that self-powered direct X-ray detectors,fabricated starting from a FAPbBr_(3)submicrometer-thick film deposition onto a mesoporous TiO_(2)scaffold,can withstand a 26-day uninterrupted X-ray exposure with negligible signal loss,demonstrating ultra-high operational stability and excellent repeatability.No structural modification is observed after irradiation with a total ionizing dose of almost 200 Gy,revealing an unexpectedly high radiation hardness for a metal-halide perovskite thin film.In addition,trap-assisted photoconductive gain enabled the device to achieve a record bulk sensitivity of 7.28 C Gy^(−1)cm^(−3)at 0 V,an unprecedented value in the field of thin-film-based photoconductors and photodiodes for“hard”X-rays.Finally,prototypal validation under the X-ray beam produced by a medical linear accelerator for cancer treatment is also introduced.

Countermeasure against blinding attack for single-photon detectors in quantum key distribution

Quantum key distribution(QKD),rooted in quantum mechanics,offers information-theoretic security.However,practi-cal systems open security threats due to imperfections,notably bright-light blinding attacks targeting single-photon detectors.Here,we propose a concise,robust defense strategy for protecting single-photon detectors in QKD systems against blinding attacks.Our strategy uses a dual approach:detecting the bias current of the avalanche photodiode(APD)to defend against con-tinuous-wave blinding attacks,and monitoring the avalanche amplitude to protect against pulsed blinding attacks.By integrat-ing these two branches,the proposed solution effectively identifies and mitigates a wide range of bright light injection attempts,significantly enhancing the resilience of QKD systems against various bright-light blinding attacks.This method forti-fies the safeguards of quantum communications and offers a crucial contribution to the field of quantum information security.

High-speed performance self-powered short wave ultraviolet radiation detectors based onκ(ε)-Ga_(2)O_(3)

High-speed solar-blind short wavelength ultraviolet radiation detectors based onκ(ε)-Ga_(2)O_(3)layers with Pt contacts were demonstrated and their properties were studied in detail.Theκ(ε)-Ga_(2)O_(3)layers were deposited by the halide vapor phase epitaxy on patterned GaN templates with sapphire substrates.The spectral dependencies of the photoelectric properties of struc-tures were analyzed in the wavelength interval 200-370 nm.The maximum photo to dark current ratio,responsivity,detectiv-ity and external quantum efficiency of structures were determined as:180.86 arb.un.,3.57 A/W,1.78×10^(12) Hz^(0.5)∙cm·W^(-1) and 2193.6%,respectively,at a wavelength of 200 nm and an applied voltage of 1 V.The enhancement of the photoresponse was caused by the decrease in the Schottky barrier at the Pt/κ(ε)-Ga_(2)O_(3)interface under ultraviolet exposure.The detectors demon-strated could functionalize in self-powered mode due to built-in electric field at the Pt/κ(ε)-Ga_(2)O_(3)interface.The responsivity and external quantum efficiency of the structures at a wavelength of 254 nm and zero applied voltage were 0.9 mA/W and 0.46%,respectively.The rise and decay times in self-powered mode did not exceed 100 ms.

基于InVEST和Geodetector模型的合肥生境质量时空演变及影响因素分析

依据2000年、2010年和2020年3期土地利用数据,运用耦合模型探究合肥2000—2020年间生境质量时空变化特征并对影响因子进行分析。结果表明:2000—2020年合肥生境质量指数平均值分别为0.185、0.181、0.178,呈逐年下降趋势,较低质量区域向低质量区域的转移面积最大。合肥生境质量退化度在2000—2020年逐年下降,呈现出“环状增长”的空间分布特征。高退化度区域主要分布在低值区域外围,呈现非连续性扩张。区域生产总值、坡度以及高程是主要影响因子,双因子增强效应显著,年末总人口和人口密度为主导单因子。

Determination of Benzo[a]pyrene in Edible Oil by High Performance Liquid Chromatography-Fluorescence Detector (HPLC-FL)

In this study, an optimized high performance liquid chromatography-fluorescence detector (HPLC-FL) method for the determination of benzo[a]pyrene in edible oil was established. HPLC was performed with Thermo Fisher Scientific C18 column (250 mm×4.6 mm, 5 μm) as the chromatographic column and acetonitrile and water as the mobile phase, and the excitation wavelength and emission wavelength of fluorescence detector were 286 and 430 nm, respectively. The response was high, and the linear range was 0.5-10.0 ng/ml. The lowest limit of detection was 0.11 ng/ml, and the average recovery was 92.5%. This method is suitable for quantitative analysis of benzo[a]pyrene content in edible oil.

一种Anchor-Free的联合模型车辆多目标跟踪算法

针对现有车辆跟踪算法不能很好地平衡效率与精度,存在较多ID切换、误检的问题,提出一种Anchor-Free的联合模型车辆多目标跟踪算法,通过把Anchor-Free可变卷积深度特征融合网络引入联合模型,在联合学习车辆目标检测和重识别特征嵌入的基础上,以轨迹关联的方式,完成车辆多目标跟踪任务.所提出的算法在UA-DETRAC车辆数据集及KITTI-tracking数据集组合成的联合数据集上进行训练和测试,结果表明,提出的车辆多目标跟踪算法有效减少了ID切换、误检问题的出现,网络结构更简单,算法运行效率更高.