基于改进Oriented R-CNN的旋转框麦穗检测与计数模型

为对干扰、遮挡等复杂的田野环境中麦穗进行精准定位与计数,该研究提出了一种改进的Oriented R-CNN麦穗旋转框检测与计数方法,首先在主干网络中引入跨阶段局部空间金字塔(spatial pyramid pooling cross stage partial networks,SPPCSPC)模块扩大模型感受野,增强网络感知能力;其次,在颈网络中结合路径聚合网络(PANet,path aggregation network)和混合注意力机制(E2CBAM,efficient two convolutional block attention module),丰富特征图包含的特征信息;最后采用柔性非极大值抑制算法(Soft-NMS,soft-non maximum suppression)优化预测框筛选过程。试验结果显示,改进的模型对复杂环境中的麦穗检测效果良好。相较原模型,平均精确度均值mAP提高了2.02个百分点,与主流的旋转目标检测模型Gliding vertex、R3det、Rotated Faster R-CNN、S2anet和Rotated Retinanet相比,mAP分别提高了4.99、2.49、3.94、2.25和4.12个百分点。该研究方法利用旋转框准确定位麦穗位置,使得框内背景区域面积大幅度减少,为实际观察麦穗生长状况和统计数量提供了一种有效的方法。

一种改进Oriented RepPoints的遥感图像有向目标检测

为解决Oriented RepPoints算法在遥感图像有向目标检测中因遥感图像背景干扰信息较多和目标尺度大小不一所导致的检测精度不高、易漏检误检等问题,提出一种改进Oriented RepPoints的遥感图像有向目标检测方法MA-RPDet(Mixed Attention RepPoints Detector).首先,采用了PVTv2作为主干网络,该网络利用线性空间缩减自注意力机制提取出更具局部连续性的特征图,并保持与卷积运算类似的线性复杂度.其次,在特征融合阶段设计了串联性混合注意力模块,进一步强化了重要特征,促进了多尺度特征的高效交互.最后,引入平滑GIoU损失函数对模型学习策略进行优化,提高了检测精度.在两个遥感图像目标检测数据集DOTA和HRSC2016上的实验结果表明,所提方法的检测精度mAP分别达到了77.19%和90.3%,均高于其他对比算法,证明了本文方法的有效性.

Three-dimensionally oriented organization of hexagonal MIL-96 microplates toward superior film microstructure

Preferential orientation control of metal—organic framework(MOF)films is advantageous for maximizing pore uniformity and minimizing grain-boundary defects.Nonetheless,the preparation of MOF films with both in-plane and out-of-plane orientations remains a grand challenge.In this study,we reported the preparation of three-dimensionally oriented MIL-96 layers through combining morphology control of MIL-96 seeds with addition of polyvinylpyrrolidone surfactants and arachidonic acids.The three-dimensionally oriented MIL-96 film was readily obtained through in-plane epitaxial growth.It is anticipated that the aforementioned protocol can be effective for obtaining diverse MOF films with a three-dimensionally oriented organization.

Oriented Bounding Box Object Detection Model Based on Improved YOLOv8

In the study of oriented bounding boxes(OBB)object detection in high-resolution remote sensing images,the problem of missed and wrong detection of small targets occurs because the targets are too small and have different orientations.Existing OBB object detection for remote sensing images,although making good progress,mainly focuses on directional modeling,while less consideration is given to the size of the object as well as the problem of missed detection.In this study,a method based on improved YOLOv8 was proposed for detecting oriented objects in remote sensing images,which can improve the detection precision of oriented objects in remote sensing images.Firstly,the ResCBAMG module was innovatively designed,which could better extract channel and spatial correlation information.Secondly,the innovative top-down feature fusion layer network structure was proposed in conjunction with the Efficient Channel Attention(ECA)attention module,which helped to capture inter-local cross-channel interaction information appropriately.Finally,we introduced an innovative ResCBAMG module between the different C2f modules and detection heads of the bottom-up feature fusion layer.This innovative structure helped the model to better focus on the target area.The precision and robustness of oriented target detection were also improved.Experimental results on the DOTA-v1.5 dataset showed that the detection Precision,mAP@0.5,and mAP@0.5:0.95 metrics of the improved model are better compared to the original model.This improvement is effective in detecting small targets and complex scenes.

Estimation of cancer cell migration in biomimetic random/oriented collagen fiber microenvironments

Increasing data indicate that cancer cell migration is regulated by extracellular matrixes and their surrounding biochemical microenvironment,playing a crucial role in pathological processes such as tumor invasion and metastasis.However,conventional two-dimensional cell culture and animal models have limitations in studying the influence of tumor microenvironment on cancer cell migration.Fortunately,the further development of microfluidic technology has provided solutions for the study of such questions.We utilize microfluidic chip to build a random collagen fiber microenvironment(RFM)model and an oriented collagen fiber microenvironment(OFM)model that resemble early stage and late stage breast cancer microenvironments,respectively.By combining cell culture,biochemical concentration gradient construction,and microscopic imaging techniques,we investigate the impact of different collagen fiber biochemical microenvironments on the migration of breast cancer MDA-MB-231-RFP cells.The results show that MDA-MB-231-RFP cells migrate further in the OFM model compared to the RFM model,with significant differences observed.Furthermore,we establish concentration gradients of the anticancer drug paclitaxel in both the RFM and OFM models and find that paclitaxel significantly inhibits the migration of MDA-MB-231-RFP cells in the RFM model,with stronger inhibition on the high concentration side compared to the low concentration side.However,the inhibitory effect of paclitaxel on the migration of MDA-MB-231-RFP cells in the OFM model is weak.These findings suggest that the oriented collagen fiber microenvironment resembling the late-stage tumor microenvironment is more favorable for cancer cell migration and that the effectiveness of anticancer drugs is diminished.The RFM and OFM models constructed in this study not only provide a platform for studying the mechanism of cancer development,but also serve as a tool for the initial measurement of drug screening.

Analysis of the Effect of Problem-Oriented Nursing Intervention on Patients with Lower Extremity Arteriosclerosis Obliterans in Vascular Surgery

Objective: To analyze the effect of problem-oriented nursing intervention on patients with lower extremity arteriosclerosis obliterans (ASO) in vascular surgery. Methods: The clinical data of 128 patients with lower extremity ASO in vascular surgery were selected and randomly divided into groups A and B, with 64 cases each. Group A is the control group, and Group B is the observation group. Group A received the routine nursing intervention, and Group B received the problem-oriented nursing intervention. The compliance, self-care ability, psychological state, quality of life, and nursing satisfaction of the two groups of patients were evaluated based on various indicators. Results: After the intervention, the evaluation of self-care ability (ESCA) score of the patients in Group B was higher than that of Group A, and the symptom checklist-90 (SCL-90) score was lower than that of Group A. The differences were significant (t = 10.019, t = 3.118, P < 0.01). After the intervention, the World Health Organization Quality of Life Brief (WHOQOL-BREF) index scores of the two groups increased and the increase in Group B was significantly higher than Group A (P < 0.001). The compliance rate of Group B (62/ 96.88%) was higher than that of Group A (52/ 81.25%), and the difference was extremely significant (χ2 = 8.020, P < 0.01). Conclusion: Problem-oriented nursing intervention for patients with lower extremity ASO in vascular surgery improved the patient’s self-care ability, and quality of life, reduced the patient’s negative emotions, and enhanced their overall satisfaction.

Intervention Effect of Lower Limb Rehabilitation Robot with Task-Oriented Training on Stroke Patients and Its Influence on KFAROM Score

Objective: To explore the effect of lower limb rehabilitation robot combined with task-oriented training on stroke patients and its influence on KFAROM score. Methods: 100 stroke patients with hemiplegia admitted to our hospital from January 2023 to December 2023 were randomly divided into two groups, the control group (50 cases) was given task-oriented training assisted by nurses, and the observation group (50 cases) was given lower limb rehabilitation robot with task-oriented training. Lower limb balance, lower limb muscle strength, motor function, ankle function, knee flexion range of motion and walking ability were observed. Results: After treatment, the scores of BBS, quadriceps femoris and hamstrings in the observation group were significantly higher than those in the control group (P Conclusion: In the clinical treatment of stroke patients, the combination of task-oriented training and lower limb rehabilitation robot can effectively improve the lower limb muscle strength, facilitate the recovery of balance function, and have a significant effect on the recovery of motor function, which can improve the walking ability of stroke patients and the range of motion of knee flexion, and achieve more ideal therapeutic effectiveness.

An Approach to Detect Structural Development Defects in Object-Oriented Programs

Structural development defects essentially refer to code structure that violates object-oriented design principles. They make program maintenance challenging and deteriorate software quality over time. Various detection approaches, ranging from traditional heuristic algorithms to machine learning methods, are used to identify these defects. Ensemble learning methods have strengthened the detection of these defects. However, existing approaches do not simultaneously exploit the capabilities of extracting relevant features from pre-trained models and the performance of neural networks for the classification task. Therefore, our goal has been to design a model that combines a pre-trained model to extract relevant features from code excerpts through transfer learning and a bagging method with a base estimator, a dense neural network, for defect classification. To achieve this, we composed multiple samples of the same size with replacements from the imbalanced dataset MLCQ1. For all the samples, we used the CodeT5-small variant to extract features and trained a bagging method with the neural network Roberta Classification Head to classify defects based on these features. We then compared this model to RandomForest, one of the ensemble methods that yields good results. Our experiments showed that the number of base estimators to use for bagging depends on the defect to be detected. Next, we observed that it was not necessary to use a data balancing technique with our model when the imbalance rate was 23%. Finally, for blob detection, RandomForest had a median MCC value of 0.36 compared to 0.12 for our method. However, our method was predominant in Long Method detection with a median MCC value of 0.53 compared to 0.42 for RandomForest. These results suggest that the performance of ensemble methods in detecting structural development defects is dependent on specific defects.

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。

基于改进Faster R-CNN的热轧带钢表面缺陷检测

热轧带钢是钢铁行业的重要产品,其表面缺陷是影响产品质量的重要因素。针对传统缺陷检测算法存在的过程繁琐、精度不足和效率低下等问题,提出一种基于改进更快速区域卷积神经网络(faster region-based convolutional neural network,Faster R-CNN)的检测算法,实现对热轧带钢表面缺陷的高效、高精度检测。首先,采用特征相加的方法对底层细节特征和高层语义特征进行融合;然后,采用精准的感兴趣区域池化(precise region of interest pooling,Precise ROI Pooling)获取固定大小的特征向量,避免特征出现位置偏差;最后,利用均值偏移聚类算法对带钢数据集进行聚类,获得适用于热轧带钢表面缺陷检测的先验框尺寸。实验结果表明,所提算法在热轧带钢表面缺陷检测数据集上的平均精度均值达到了85.34%,检测速度为23.5帧/s,且鲁棒性良好,满足实际的工业检测需求。

基于空间注意力机制的Mask R-CNN致密储层岩石薄片图像鉴定

针对陆相致密储层岩石薄片鉴定识别难、制片成本高、时间消耗长和人为主观强等难题,选取鄂尔多斯盆地临兴区块上古生界和松辽盆地三肇凹陷扶余油层为靶区,提出一种基于深度学习的致密油储层岩石薄片人工智能鉴定方法,引入图像预处理技术去除岩石薄片图像噪声并统一图像像素大小,构建空间几何增广机制,基于空间注意力机制改进Mask R-CNN算法,并将上述方法应用于实例靶区进行有效性验证。结果表明:图像预处理技术能够在保障图像特征的前提下,有效提高图像质量,减少噪声干扰;空间几何图像增广机制能够在在一定程度上增加可用样本的数量;基于空间注意力机制的Mask R-CNN算法可以同时完成复杂岩石薄片成分的分割与智能识别工作,分割精度在不同数据集情况下的平均精度为89.2%,整体识别准确率为93%,适用于致密油储层岩石薄片特征鉴定。

改进Mask R-CNN的馆藏报纸图像内容分割

开展馆藏报纸图像内容分割研究,能提升文字识别准确率,对促进机器识别取代人工操作、提高图书馆数字化工作效率具有重要意义。文章根据报纸图像呈现的特征,提出一种基于改进MaskR-CNN的算法,实现报纸图像内容分割。首先,通过优化锚框比例和损失函数,对原始MaskR-CNN算法进行改进。其次,采用数据增强、调整训练参数开展样本训练。最后,通过实验的方式对改进后的MaskR-CNN算法训练模型和原始算法训练模型进行比较,并采用AP_bbox和AP_segm评价指标对实验结果进行评估,改进后的算法训练模型AP_bbox为0.935,AP_segm为0.943,均超过原始算法训练模型。实验结果表明,改进后的MaskR-CNN算法能够实现报纸图像内容有效检测与分割。

基于改进Mask R-CNN的青菜杂质检测研究

绿叶蔬菜的智能包装加工是实现绿叶蔬菜智能化生产、降低生产成本的重要部分,对绿叶蔬菜在包装加工时的杂质检测是其重要前提。以青菜为研究对象,提出一种基于Mask R-CNN的青菜杂质检测模型。首先采集标注掺杂枯树叶、枯菜叶和碎纸片3种常见杂质的青菜图像1370多张,并通过数据增强的方法扩充建立含有2740张青菜杂质图像的数据集。为减少背景对杂质检测的影响,通过在Mask R-CNN模型中加入协调注意力机制,同时添加全连接层和Dropout层,增强模型特征提取能力,减少过拟合现象,并使用迁移学习方法对模型进行微调。结果表明改进后的Mask R-CNN算法对青菜杂质识别的平均精度均值为99.19%,检测速度为8.45 FPS,检测效果良好,可以满足青菜杂质的检测需求。

基于改进Faster R-CNN与U-Net算法的桥梁病害识别与量化方法

为实现桥梁病害检测的自动化,对基于图像处理技术的混凝土桥梁表观病害的智能识别和尺寸确定方法展开研究.提出基于改进Faster R-CNN算法的病害识别方法,利用K均值聚类和遗传算法对区域候选网络锚框进行优化设计;以裂缝预测区域为基础,提出ResNet34结合U-Net的裂缝形态提取方法,并结合裂缝形态学研究了裂缝像素宽度和长度的确定方法.结果表明:锚框优化设计可改进Faster R-CNN算法的表观病害识别效果,5类常见病害的预测准确率、召回率、平均精确率分别由68.40%、69.87%、74.64%提升到85.40%、83.59%、83.72%;利用病害预测框,结合改进U-Net算法的裂缝像素尺寸计算,可实现裂缝病害尺寸的自动测量;基于改进Faster R-CNN和改进U-Net的方法可实现混凝土桥梁常见病害的智能识别和尺寸量化,从而提高桥梁病害检测效率并促进桥梁技术状况评定的智能化.

基于FPGA加速的Mask R-CNN稻瘟病高通量自适应识别模型研究

针对基于图像的稻瘟病现场检测技术依赖先验知识且受制于算力与田间网络状况,无法实现自适应实时检测的问题,提出一种可利用现场可编程门阵列(Field programmable gate array,FPGA)加速的Mask R-CNN(Mask region-based convolutional neural network)稻瘟病高通量自适应快速识别模型。首先将骨干网络改进为MobileNetV2,利用其倒残差模块降低计算量,提高模型并行处理能力;随后增加用于稻瘟病多尺度特征融合的特征金字塔网络模块,使模型具备多尺度自适应处理能力;最后由全卷积网络(Fully convolutional network,FCN)分支输出稻瘟病病斑的实例分割,同时使用交叉熵损失函数完成稻瘟病的定位与分类。稻瘟病实测数据集对模型的验证结果表明:当输入为全高清图像时,模型平均推理时间减少至85 ms,相较GPU服务器、同级别GPU边缘计算平台,速度分别提高86.2%、63.0%。在交并比为0.6时,准确率可达98.0%,病斑捕获能力平均提升21.2%。提出的Mask R-CNN自适应快速识别模型能够在田间恶劣网络状况下实现稻瘟病的快速现场检测,具有更好的抗噪能力和鲁棒性能,为水稻病害实时检测、察打一体提供了高效实时的片上系统方案。

基于改进Mask R-CNN的笼养死鸭识别方法

针对规模化笼养肉鸭舍内死鸭识别采用人工作业方式时,存在作业效率低、劳动强度大、养殖成本高等问题,以层叠式笼养肉鸭为研究对象,提出了一种基于深度学习的笼养死鸭识别方法。为了采集数据,首先面向立体层叠式养殖环境设计了一款适用于肉鸭舍的自主巡检装备。针对笼养肉鸭舍铁丝网遮挡严重的问题,基于机器视觉对笼网进行修复,基于OpenCV对图像进行增强处理。构建了一种基于Mask R-CNN的死鸭识别模型,采用Swin Transformer对模型进行优化,解决了Mask R-CNN网络缺乏整合全局信息能力的问题。对比分析了SOLO v2、Mask R-CNN和Mask R-CNN+Swin Transformer模型识别笼内死鸭准确率。实验结果表明,在平均精度均值为90%的条件下,Mask R-CNN+Swin Transformer模型对笼内死鸭总体识别准确率可达95.8%,在自主巡检装备上的检测效果优于其他主流的目标检测算法。

基于Mask R-CNN的电力关键设备运行状态检测

为判断电力关键设备两端负载电压与干路总电压之间的数值关系,提出基于Mask R-CNN的电力关键设备运行状态检测方法。在Mask R-CNN网络结构中,判断电量信号的时域范围与频域范围。然后根据负荷阻抗特征计算交流参数的取值结果,再联合已获取的电信号参量,求解连续相关函数,从而检测电力关键设备运行状态。实验过程中,设备两端负载电压、内阻消耗电压之和与干路总电压之间的差值未超过1.5 V,说明该方法能够证明电力关键设备两端负载电压、内阻消耗电压之和等于干路总电压数值的猜想成立,可以根据该验证检测电力关键设备运行状态是否正常。

优化Faster R-CNN算法的小样本缺陷检测研究

随着自动化检测技术的发展,基于深度学习的缺陷检测技术以其高精度、高效率、非接触性的特点正逐渐成为工业和学术领域的研究热点。为解决实际工业生产中由于产品缺陷数据集样本不足、类别不均衡导致的模型过拟合、检测精度低等问题,提出了一种基于Faster R-CNN算法框架优化的缺陷检测模型Faster R-CNN-H-BFC,通过基于多层感知器(multi-layer perceptron,MLP)实现的幻觉网络能够从具有丰富样本的基类中学习到类共享特征并为新类生成额外的幻觉样本供模型训练,并且针对Faster R-CNN本身存在的识别精度低以及检测效果差等问题,将原始的VGG16主干网络替换为具有残差结构的ResNet50,并引入了特征金字塔网络(Feature Pyramid Networks,FPN)实现多尺度特征融合,添加混合注意力机制(Convolutional Block Attention Module,CBAM)来增强模型的特征提取能力。实验和数据表明:改进后的缺陷检测模型在极少样本场景下具有较好的检测效果,平均检测精度相较于改进前提升了3.11%。

基于改进Faster R-CNN的变电站设备外部缺陷检测

针对变电站设备外部缺陷目标检测任务中目标形状多样,周围环境复杂,当前代表性算法识别准确度低,错检漏检严重的问题,对比了众多目标检测算法在变电站设备缺陷数据集上的检测结果,检测精度较高的是添加了特征融合金字塔结构的Faster R-CNN(faster region-based convolutional network)算法,但其对小目标物体和设备渗漏油的检测精度仍有提升空间,为此设计一种基于Faster R-CNN的改进算法。改进算法通过对输入图像进行数据增强,在网络中添加SPP(spatial pyramid pooling)结构以及改进特征融合方式,对分类以及边界框回归损失函数进行改进的方式来提高缺陷的检测精度。与原Faster R-CNN算法进行对比,改进算法在变电站设备缺陷目标检测数据集的检测结果中AP(average precision)(0.5∶0.95)提高了2.7个百分点,AP(0.5)提高了4.3个百分点,对小目标物体的检测精度也提高了1.8个百分点,试验结果验证了该方法的有效性。

基于Faster R-CNN的动漫场景多人物自动识别研究

当前动漫场景多人物识别方法在提取目标特征时,对于汇聚特征信息的滑动窗口定位不准确,所提取到的特征信息不准确,导致识别精度较差,因此为了解决这一问题,提出了一种基于Faster R-CNN的动漫场景多人物自动识别方法。标注大量动漫场景人物形象图片,构建训练集和测试集,将其输入Faster R-CNN神经网络模型,提取图像特征并构建特征图。采用滑动窗口遍历特征图,选择特征向量评分最高的窗口,保证窗口内局部特征可以充分表示动漫人物主要特征,根据特征提取结果自动识别多个动漫人物身份,完成动漫场景多人物自动识别。实验结果表明,设计方法与两种传统方法相比,人物识别召回率分别提升了11.10%和18.99%,提高了目标识别精度,人物识别过拟合比率稳定在1.0060,说明该方法能够高精度不同类别的动漫人物,且识别过程较为稳定,识别效率较高。