Control of GaN inverted pyramids growth on c-plane patterned sapphire substrates

Growth of gallium nitride(GaN)inverted pyramids on c-plane sapphire substrates is benefit for fabricating novel devices as it forms the semipolar facets.In this work,GaN inverted pyramids are directly grown on c-plane patterned sapphire substrates(PSS)by metal organic vapor phase epitaxy(MOVPE).The influences of growth conditions on the surface morphol-ogy are experimentally studied and explained by Wulff constructions.The competition of growth rate among{0001},{1011},and{1122}facets results in the various surface morphologies of GaN.A higher growth temperature of 985 ℃ and a lowerⅤ/Ⅲratio of 25 can expand the area of{}facets in GaN inverted pyramids.On the other hand,GaN inverted pyramids with almost pure{}facets are obtained by using a lower growth temperature of 930℃,a higherⅤ/Ⅲratio of 100,and PSS with pattern arrangement perpendicular to the substrate primary flat.

An anti-aliasing filtering of quantum images in spatial domain using a pyramid structure

As a part of quantum image processing,quantum image filtering is a crucial technology in the development of quantum computing.Low-pass filtering can effectively achieve anti-aliasing effects on images.Currently,most quantum image filterings are based on classical domains and grayscale images,and there are relatively fewer studies on anti-aliasing in the quantum domain.This paper proposes a scheme for anti-aliasing filtering based on quantum grayscale and color image scaling in the spatial domain.It achieves the effect of anti-aliasing filtering on quantum images during the scaling process.First,we use the novel enhanced quantum representation(NEQR)and the improved quantum representation of color images(INCQI)to represent classical images.Since aliasing phenomena are more pronounced when images are scaled down,this paper focuses only on the anti-aliasing effects in the case of reduction.Subsequently,we perform anti-aliasing filtering on the quantum representation of the original image and then use bilinear interpolation to scale down the image,achieving the anti-aliasing effect.The constructed pyramid model is then used to select an appropriate image for upscaling to the original image size.Finally,the complexity of the circuit is analyzed.Compared to the images experiencing aliasing effects solely due to scaling,applying anti-aliasing filtering to the images results in smoother and clearer outputs.Additionally,the anti-aliasing filtering allows for manual intervention to select the desired level of image smoothness.

IMTNet:Improved Multi-Task Copy-Move Forgery Detection Network with Feature Decoupling and Multi-Feature Pyramid

Copy-Move Forgery Detection(CMFD)is a technique that is designed to identify image tampering and locate suspicious areas.However,the practicality of the CMFD is impeded by the scarcity of datasets,inadequate quality and quantity,and a narrow range of applicable tasks.These limitations significantly restrict the capacity and applicability of CMFD.To overcome the limitations of existing methods,a novel solution called IMTNet is proposed for CMFD by employing a feature decoupling approach.Firstly,this study formulates the objective task and network relationship as an optimization problem using transfer learning.Furthermore,it thoroughly discusses and analyzes the relationship between CMFD and deep network architecture by employing ResNet-50 during the optimization solving phase.Secondly,a quantitative comparison between fine-tuning and feature decoupling is conducted to evaluate the degree of similarity between the image classification and CMFD domains by the enhanced ResNet-50.Finally,suspicious regions are localized using a feature pyramid network with bottom-up path augmentation.Experimental results demonstrate that IMTNet achieves faster convergence,shorter training times,and favorable generalization performance compared to existingmethods.Moreover,it is shown that IMTNet significantly outperforms fine-tuning based approaches in terms of accuracy and F_(1).

Two-Layer Attention Feature Pyramid Network for Small Object Detection

Effective small object detection is crucial in various applications including urban intelligent transportation and pedestrian detection.However,small objects are difficult to detect accurately because they contain less information.Many current methods,particularly those based on Feature Pyramid Network(FPN),address this challenge by leveraging multi-scale feature fusion.However,existing FPN-based methods often suffer from inadequate feature fusion due to varying resolutions across different layers,leading to suboptimal small object detection.To address this problem,we propose the Two-layerAttention Feature Pyramid Network(TA-FPN),featuring two key modules:the Two-layer Attention Module(TAM)and the Small Object Detail Enhancement Module(SODEM).TAM uses the attention module to make the network more focused on the semantic information of the object and fuse it to the lower layer,so that each layer contains similar semantic information,to alleviate the problem of small object information being submerged due to semantic gaps between different layers.At the same time,SODEM is introduced to strengthen the local features of the object,suppress background noise,enhance the information details of the small object,and fuse the enhanced features to other feature layers to ensure that each layer is rich in small object information,to improve small object detection accuracy.Our extensive experiments on challenging datasets such as Microsoft Common Objects inContext(MSCOCO)and Pattern Analysis Statistical Modelling and Computational Learning,Visual Object Classes(PASCAL VOC)demonstrate the validity of the proposedmethod.Experimental results show a significant improvement in small object detection accuracy compared to state-of-theart detectors.

Loading direction dependence of asymmetric response of pyramidal slip in rolled AZ31 magnesium alloy

Textured magnesium alloys usually exhibit anisotropic mechanical behavior due to the asymmetric activation of different twinning and slipping modes.This work focuses on the pyramidal slip responses of rolled AZ31 magnesium alloy under two loading conditions,compressive and tensile loading along the normal direction.Under the condition where the compressive loading direction is closely parallel to the c-axis of the unit cell,tensile twinning and basal slips are prohibited, dislocations then active and tend to accumulate at grain boundaries and form dislocation walls.Meanwhile,these dislocations exhibit zigzag morphologies,which result from the cross-slip from {10■1} first-order pyramidal plane to {11■2} second-order pyramidal plane,then back to {10■1} first-order pyramidal plane.Under the condition where tensile twins are prevalent,{10■1} first-order and {11■2} second-order pyramidal dislocations are favorable to be activated.Both types of dislocations behave climb-like dissociations onto the basal plane,forming zigzag dislocations.

Pyramiding of multiple genes generates rapeseed introgression lines with clubroot and herbicide resistance, high oleic acid content, and early maturity

Clubroot and herbicide resistance,high oleic acid(OA)content,and early maturity are targets of rapeseed(Brassica napus L.)breeding.The objective of this study was to develop new male-fertility restorer lines by pyramiding favorable genes to improve these traits simultaneously.Seven elite alleles for the four traits were introduced into the restorer line 621R by speed breeding with marker-assisted and phenotypic selection.Six introgression lines(ILs)were developed with four-to seven-gene combinations and crossed with two elite parents to develop hybrids.All ILs and their corresponding hybrids displayed high resistance to both clubroot pathotype 4 and sulfonylurea herbicides.Three ILs and their hybrids showed large increases in OA contents and four showed earlier maturity.These new ILs may be useful in rapeseed hybrid breeding for the target traits.

Semantic segmentation of pyramidal neuron skeletons using geometric deep learning

Neurons can be abstractly represented as skeletons due to the filament nature of neurites.With the rapid development of imaging and image analysis techniques,an increasing amount of neuron skeleton data is being produced.In some scienti fic studies,it is necessary to dissect the axons and dendrites,which is typically done manually and is both tedious and time-consuming.To automate this process,we have developed a method that relies solely on neuronal skeletons using Geometric Deep Learning(GDL).We demonstrate the effectiveness of this method using pyramidal neurons in mammalian brains,and the results are promising for its application in neuroscience studies.

Fine-Grained Action Recognition Based on Temporal Pyramid Excitation Network

Mining more discriminative temporal features to enrich temporal context representation is considered the key to fine-grained action recog-nition.Previous action recognition methods utilize a fixed spatiotemporal window to learn local video representation.However,these methods failed to capture complex motion patterns due to their limited receptive field.To solve the above problems,this paper proposes a lightweight Temporal Pyramid Excitation(TPE)module to capture the short,medium,and long-term temporal context.In this method,Temporal Pyramid(TP)module can effectively expand the temporal receptive field of the network by using the multi-temporal kernel decomposition without significantly increasing the computational cost.In addition,the Multi Excitation module can emphasize temporal importance to enhance the temporal feature representation learning.TPE can be integrated into ResNet50,and building a compact video learning framework-TPENet.Extensive validation experiments on several challenging benchmark(Something-Something V1,Something-Something V2,UCF-101,and HMDB51)datasets demonstrate that our method achieves a preferable balance between computation and accuracy.

Potential Power of the Pyramidal Structure VII: Effects of Pyramid Power and Bio-Entanglement on the Circadian Rhythm of Biosensors

We have demonstrated the existence of a pyramid power and have revealed its characteristics by strictly scientific experiments using biosensors. We also revealed the existence of a Bio-Entanglement, an entangled relationship between biosensors. A parallel study of biosensors (edible cucumber slices) had also been conducted, and we found that the circadian rhythm of gas concentrations emitted from biosensors changes seasonally. The pyramid power and Bio-Entanglement did not change the number of cycles in the periodic approximation curve representing circadian rhythm. Therefore, in this paper we analyzed the influence of the pyramid power and Bio-Entanglement, i.e., their influence on the phase, amplitude, and correlation coefficient of the periodic approximation curve representing the circadian rhythm of emitted gas concentrations. The main results are as follows. 1) The pyramid power shifted the phase of the periodic approximation curve representing the circadian rhythm by 43 minutes. 2) The amplitude of the periodic approximation curve changed with the pyramid power and the Bio-Entanglement. The effect on the lower and upper sections of the biosensors stacked in two layers was different, with a tendency to increase the amplitude of the lower layer and decrease the amplitude of the upper layer. 3) The pyramid power and the Bio-Entanglement affected the correlation coefficient between gas concentration and the periodic approximation curve representing the circadian rhythm of gas concentration. The effect on the lower and upper layers of the biosensors was different, with a tendency for the lower layer correlation coefficient to be larger and the upper layer correlation coefficient to be smaller. Previously we demonstrated that the pyramid power and the Bio-Entanglement affect the ratio of gas concentration, i.e., psi index Ψ. In this paper we demonstrate for the first time that the pyramid power and the Bio-Entanglement affect time, i.e., phase difference.

Dissociation of edge and screw pyramidal Ⅰ and Ⅱ dislocations in magnesium

Pyramidal dislocations in magnesium (Mg) and other hexagonal close-packed metals play an important role in accommodating plastic strains along the c-axis.Bulk single crystal Mg only presents very limited plasticity in c-axis compression,and this behavior was attributed to out-of-plane dissociation of pyramidal dislocations onto the basal plane and resulted in an immobile dislocation configuration.In contrast,other simulations and experiments reported in-plane dissociation of pyramidal dislocations on their slip planes.Thus,the core structure and mode of dissociation of pyramidal dislocations are still not well understood.To better understand the dissociation behavior of pyramidal dislocations in Mg at room temperature,in this work,atomistic simulations were conducted to investigate four types of pyramidal dislocations at 300 K:edge and screw Py-Ⅰ on{1011},edge and screw Py-Ⅱ on{1122}by using a modified embedded atom method (MEAM) potential for Mg and anisotropic elasticity dislocation model.The results show that when energy minimization was performed before relaxation,in-plane dissociation of edge dislocations on respective pyramidal plane could be obtained at room temperature for all four types of dislocation.Without energy minimization,the edge dislocations dissociated out-of-plane onto the basal plane.Calculations of potential energy and hydrostatic stress of individual atoms at the edge dislocation core show that the extraordinarily high energy and atomic stresses in the as-constructed dislocation structures caused the out-of-plane dissociation onto the basal plane.The core structures of all four types of pyramidal dislocation after in-plane dissociation were analyzed by computing the distribution of the Burgers vector.

Potential Power of the Pyramidal Structure VIII: Exploration of Periodic Diurnal Oscillation of Pyramid Power and Bio-Entanglement

To date, numerous books have been published on so-called “pyramid power” but there have been few academic papers on this subject other than our own. Since 2007, to demonstrate the pyramid power, we have undertaken strictly scientific experiments using a pyramidal structure (PS) that we have carefully constructed. In previous reports, we used the edible cucumber, Cucumis sativus as an effective and practical biosensor. Through measurement and analysis of volatile components (gas concentrations) emitted from the biosensor, we were able to demonstrate the existence of the pyramid power and revealed some of its characteristics. In a paper published in 2022, we showed that gas concentration release from this biosensor displayed a circadian rhythm and that this rhythm changed with the season. Based on the result that the biosensor had a periodic diurnal oscillation called a circadian rhythm, we questioned whether or not pyramid power and Bio-Entanglement also had periodic diurnal oscillations. In this paper, we investigated that possibility. Our results have shown that pyramid power and Bio-Entanglement do not exhibit significant periodic diurnal oscillations. Thus we have revealed for the first time that the field associated with pyramid power is a type of static field that always exerts a constant influence. We expect that our research results will be widely accepted in the future and will become the foundation for a new research field in science, with a wide range of applications.

3D Echocardiogram Reconstruction Employing a Flip Directional Texture Pyramid

Three dimensional(3D)echocardiogram enables cardiologists to visua-lize suspicious cardiac structures in detail.In recent years,this three-dimensional echocardiogram carries important clinical value in virtual surgical simulation.However,this 3D echocardiogram involves a trade-off difficulty between accu-racy and efficient computation in clinical diagnosis.This paper presents a novel Flip Directional 3D Volume Reconstruction(FD-3DVR)method for the recon-struction of echocardiogram images.The proposed method consists of two main steps:multiplanar volumetric imaging and 3D volume reconstruction.In the crea-tion of multiplanar volumetric imaging,two-dimensional(2D)image pixels are mapped into voxels of the volumetric grid.As the obtained slices are discontin-uous,there are some missing voxels in the volume data.To restore the structural and textural information of 3D ultrasound volume,the proposed method creates a volume pyramid in parallel with theflip directional texture pyramid.Initially,the nearest neighbors of missing voxels in the multiplanar volumetric imaging are identified by 3D ANN(Approximate Nearest Neighbor)patch matching method.Furthermore,aflip directional texture pyramid is proposed and aggregated with distance in patch matching tofind out the most similar neighbors.In the recon-struction step,structural and textural information obtained from differentflip angle directions can reconstruct 3D volume well with the desired accuracy.Com-pared with existing 3D reconstruction methods,the proposed Flip Directional 3D Volume Reconstruction(FD-3DVR)method provides superior performance for the mean peak signal-to-noise ratio(40.538 for the proposed method I and 39.626 for the proposed method II).Experimental results performed on the cardi-ac datasets demonstrate the efficiency of the proposed method for the reconstruc-tion of echocardiogram images.

Configuration Synthesis and Lightweight Networking of Deployable Mechanism Based on a Novel Pyramid Module

Deployable mechanism with preferable deployable performance,strong expansibility,and lightweight has attracted much attention because of their potential in aerospace.A basic deployable pyramid unit with good deployability and expandability is proposed to construct a sizeable deployable mechanism.Firstly,the basic unit folding principle and expansion method is proposed.The configuration synthesis method of adding constraint chains of spatial closed-loop mechanism is used to synthesize the basic unit.Then,the degree of freedom of the basic unit is analyzed using the screw theory and the link dismantling method.Next,the three-dimensional models of the pyramid unit,expansion unit,and array unit are established,and the folding motion simulation analysis is carried out.Based on the number of components,weight reduction rate,and deployable rate,the performance characteristics of the three types of mechanisms are described in detail.Finally,prototypes of the pyramid unit,combination unit,and expansion unit are developed to verify further the correctness of the configuration synthesis based on the pyramid.The proposed deployable mechanism provides aference for the design and application of antennas with a large aperture,high deployable rate,and lightweight.It has a good application prospect in the aerospace field.

Experimental investigation of the impact of water depth, inlet water temperature, and fins on the productivity of a Pyramid Solar Still

This experimental study aimed to investigate the impact of water depth, inlet water temperature,and fins on the productivity of a pyramid solar still in producing distilled water. The experiment was conducted in three parts, where the first part explored the variation in water depth from 1 cm to 5 cm, the second part evaluated the effect of increasing inlet water temperature from 30℃ to 50℃, and the third part added fins at the bottom of the still at a specific inlet water depth. Results showed that basin depth had a significant impact on the still's production, with a maximum variation of 40.6% observed when the water level was changed from 1 cm to 5 cm. The daily freshwater production from the pyramid solar still ranged from 3.41 kg/m~2 for a water depth of 1 cm to 2.02 kg/m~2 for a depth of 5 cm. Adding fins at the bottom of the pyramid solar still led to a 7.5% increase in productivity, while adjusting the inlet water temperature from 30℃ to 40℃ and 50℃ resulted in a 15.3% and 21.2% increase, respectively. These findings highlighted the essential factors that can influence the productivity of pyramid solar stills and can be valuable in designing and operating efficient water desalination and purification technologies.

PAN-DeSpeck:A Lightweight Pyramid and Attention-Based Network for SAR Image Despeckling

SAR images commonly suffer fromspeckle noise,posing a significant challenge in their analysis and interpretation.Existing convolutional neural network(CNN)based despeckling methods have shown great performance in removing speckle noise.However,these CNN-basedmethods have a fewlimitations.They do not decouple complex background information in amulti-resolutionmanner.Moreover,they have deep network structures thatmay result in many parameters,limiting their applicability tomobile devices.Furthermore,extracting key speckle information in the presence of complex background is also a major problem with SAR.The proposed study addresses these limitations by introducing a lightweight pyramid and attention-based despeckling(PAN-Despeck)network.The primary objective is to enhance image quality and enable improved information interpretation,particularly on mobile devices and scenarios involving complex backgrounds.The PAN-Despeck network leverages domainspecific knowledge and integrates Gaussian Laplacian image pyramid decomposition for multi-resolution image analysis.By utilizing this approach,complex background information can be effectively decoupled,leading to enhanced despeckling performance.Furthermore,the attention mechanism selectively focuses on key speckle features and facilitates complex background removal.The network incorporates recursive and residual blocks to ensure computational efficiency and accelerate training speed,making it lightweight while maintaining high performance.Through comprehensive evaluations,it is demonstrated that PAN-Despeck outperforms existing image restoration methods.With an impressive average peak signal-to-noise ratio(PSNR)of 28.355114 and a remarkable structural similarity index(SSIM)of 0.905467,it demonstrates exceptional performance in effectively reducing speckle noise in SAR images.The source code for the PAN-DeSpeck network is available on GitHub.

Deep Pyramidal Residual Network for Indoor-Outdoor Activity Recognition Based on Wearable Sensor

Recognition of human activity is one of the most exciting aspects of time-series classification,with substantial practical and theoretical impli-cations.Recent evidence indicates that activity recognition from wearable sensors is an effective technique for tracking elderly adults and children in indoor and outdoor environments.Consequently,researchers have demon-strated considerable passion for developing cutting-edge deep learning sys-tems capable of exploiting unprocessed sensor data from wearable devices and generating practical decision assistance in many contexts.This study provides a deep learning-based approach for recognizing indoor and outdoor movement utilizing an enhanced deep pyramidal residual model called Sen-PyramidNet and motion information from wearable sensors(accelerometer and gyroscope).The suggested technique develops a residual unit based on a deep pyramidal residual network and introduces the concept of a pyramidal residual unit to increase detection capability.The proposed deep learning-based model was assessed using the publicly available 19Nonsens dataset,which gathered motion signals from various indoor and outdoor activities,including practicing various body parts.The experimental findings demon-strate that the proposed approach can efficiently reuse characteristics and has achieved an identification accuracy of 96.37%for indoor and 97.25%for outdoor activity.Moreover,comparison experiments demonstrate that the SenPyramidNet surpasses other cutting-edge deep learning models in terms of accuracy and F1-score.Furthermore,this study explores the influence of several wearable sensors on indoor and outdoor action recognition ability.

Intelligent identification of oceanic eddies in remote sensing data via Dual-Pyramid UNet

海洋涡旋是大洋中重要的组成部分,对海洋能量和物质的输送至关重要.海洋涡旋的检测和表征无论是对于海洋气象学,海洋声学还是海洋生物学等领域都具有重要的研究价值.本文基于UNet架构,并结合金字塔分割注意力(PSA)模块和空洞空间卷积池化金字塔(ASPP)构造了Dual-Pyramid UNet模型,以平面异常和海表面温度数据中进行海洋涡旋的识别.实验在北大西洋和南大西洋两个涡旋活跃区域进行并选用多个评价指标对识别结果进行评价以证明模型的优异性能.

西北条锈菌源区冬小麦育种抗条锈病基因的利用现状与策略

【目的】了解中国小麦条锈病西北重要越夏菌源区甘肃陇东和陇南近20年育成小麦品种(系)中抗条锈病基因的利用情况,为陇南小麦条锈病遗传多样性控制,持久抗病品种的培育和可持续绿色健康生态农业奠定基础。【方法】于2019—2020年和2020—2021年种植季对117份冬小麦品种(系)甘肃清水县和四川郫都区进行成株期抗条锈病鉴定,2021年分别用条锈菌CYR33和CYR34对117份品种(系)进行苗期抗条锈病鉴定,并利用分子标记对15个抗条锈病基因进行检测。【结果】田间成株期对条锈菌主要流行致病性混合小种抗病的品种(系)占29.1%,其中,25.6%为陇南区域品种(系),3.4%为陇东区域品种(系),另有25.6%陇南区域感病品种(系)表现慢病性(严重度<20%)。有70.1%品种(系)苗期对CYR33表现抗病,其中,57.3%为陇南区域品种(系),12.8%为陇东区域品种(系);仅6.0%品种(系)苗期对CYR34具有抗病性,为陇南区域的兰天131等7个品种(系)。苗期和成株期抗病品种(系)以2010年后育成的兰天、中梁、天选、兰航选系列品种为主。分子标记检测结合系谱分析,发现抗条锈病基因Yr9、Yr10、Yr17、Yr18、Yr26、Yr28、Yr29、Yr30、Yr41、Yr46、YrZH22和YrZH84在所有检测品种(系)中的频率分别为49.6%、1.7%、12.8%、7.7%、12.8%、20.5%、10.3%、34.2%、2.6%、16.2%、15.4%和27.4%,且多以基因聚合体形式存在于品种(系)中,62.4%品种(系)中聚合了2—5个抗条锈病基因,YrZH84、YrZH22和Yr17中的一个或多个基因与其他抗条锈病基因聚合后,品种(系)的条锈病抗性明显高于其他基因组合。此外,在陇南菌源区,以种植小麦为主的山旱地区域(条锈菌越夏区)所推广的品种(系)中,含有Yr10、Yr17、Yr18、Yr28、Yr29、Yr30、Yr41和Yr46的频率较高,越冬区(川水地)推广品种(系)中主要有Yr26、Yr30、YrZH22和YrZH84,越夏区和越冬区品种(系)的抗性遗传背景差异明显,且越夏区品种(系)含有的抗病基因类型多样性高于越冬区。【结论】在越夏菌源区甘肃陇南、陇东的品种(系)中,抗条锈病基因分布频率、抗病基因类型及数量均有明显提高,避免了品种抗病遗传背景单一化问题,实现了品种(系)中的抗病基因较为复杂多样,部分品种的抗病性保持时间长,表明陇南地区利用小麦品种遗传多样性控制条锈病策略的实施已初显成效。

基于改进Faster R-CNN的苹果采摘视觉定位与检测方法

针对采摘机器人对场景中目标分布密集、果实相互遮挡的检测及定位能力不理想问题,提出一种引入高效通道注意力机制(ECA)和多尺度融合特征金字塔(FPN)改进Faster R-CNN果实检测及定位方法。首先,利用表达能力较强的融合FPN的残差网络ResNet50替换原VGG16网络,消除了网络退化问题,进而提取更加抽象和丰富的语义信息,提升模型对多尺度和小目标的检测能力;其次,引入注意力机制ECA模块,使特征提取网络聚焦特征图像的局部高效信息,减少无效目标的干扰,提升模型检测精度;最后,采用一种枝叶插图数据增强方法改进苹果数据集,解决图像数据不足问题。基于构建的数据集,使用遗传算法优化K-means++聚类生成自适应锚框,提高模型定位准确性。试验结果表明,改进模型对可抓取和不可直接抓取苹果的精度均值分别为96.16%和86.95%,平均精度均值为92.79%,较传统Faster R-CNN提升15.68个百分点;对可抓取和不可直接抓取的苹果定位精度分别为97.14%和88.93%,较传统Faster R-CNN分别提高12.53个百分点和40.49个百分点;内存占用量减少38.20%,每帧平均计算时间缩短40.7%,改进后的模型参数量小且实时性好,能够更好地应用于果实采摘机器人视觉系统。

结合主动光源和改进YOLOv5s模型的夜间柑橘检测方法

【目的】解决夜间环境下遮挡和较小柑橘难以准确识别的问题,实现采摘机器人全天候智能化作业。【方法】提出一种结合主动光源的夜间柑橘识别方法。首先,通过分析主动光源下颜色特征不同的夜间柑橘图像,选择最佳的光源色并进行图像采集。然后,提出一种夜间柑橘检测模型BI-YOLOv5s,该模型采用双向特征金字塔网络(Bi-FPN)进行多尺度交叉连接和加权特征融合,提高对遮挡和较小果实的识别能力;引入Coordinate attention(CA)注意力机制模块,进一步加强对目标位置信息的提取;采用融入Transformer结构的C3TR模块,在减少计算量的同时更好地提取全局信息。【结果】本文提出的BI-YOLOv5s模型在测试集上的精准率、召回率、平均准确率分别为93.4%、92.2%和97.1%,相比YOLOv5s模型分别提升了3.2、1.5和2.3个百分点。在所采用的光源色环境下,模型对夜间柑橘识别的正确率为95.3%,相比白光环境下提高了10.4个百分点。【结论】本文提出的方法对夜间环境下遮挡和小目标柑橘的识别具有较高的准确性,可为夜间果蔬智能化采摘的视觉精准识别提供技术支持。