Underwater Sea Cucumber Target Detection Based on Edge-Enhanced Scaling YOLOv4

Sea cucumber detection is widely recognized as the key to automatic culture.The underwater light environment is complex and easily obscured by mud,sand,reefs,and other underwater organisms.To date,research on sea cucumber detection has mostly concentrated on the distinction between prospective objects and the background.However,the key to proper distinction is the effective extraction of sea cucumber feature information.In this study,the edge-enhanced scaling You Only Look Once-v4(YOLOv4)(ESYv4)was proposed for sea cucumber detection.By emphasizing the target features in a way that reduced the impact of different hues and brightness values underwater on the misjudgment of sea cucumbers,a bidirectional cascade network(BDCN)was used to extract the overall edge greyscale image in the image and add up the original RGB image as the detected input.Meanwhile,the YOLOv4 model for backbone detection is scaled,and the number of parameters is reduced to 48%of the original number of parameters.Validation results of 783images indicated that the detection precision of positive sea cucumber samples reached 0.941.This improvement reflects that the algorithm is more effective to improve the edge feature information of the target.It thus contributes to the automatic multi-objective detection of underwater sea cucumbers.

A Security Enhancement Model Based on Switching Edge Strategy in Interdependent Heterogeneous Cyber-Physical Systems

With the advent of cross-domain interconnection,large-scale sensor network systems such as smart grids,smart homes,and intelligent transportation have emerged.These complex network systems often have a CPS(Cyber-Physical System)architecture and are usually composed of multiple interdependent systems.Minimal faults between interdependent networks may cause serious cascading failures between the entire system.Therefore,in this paper,we will explore the robustness detection schemes for interdependent networks.Firstly,by calculating the largest giant connected component in the entire system,the security of interdependent network systems under different attack models is analyzed.Secondly,a comparative analysis of the cascade failure mechanism between interdependent networks under the edge enhancement strategy is carried out.Finally,the simulation results verify the impact of system reliability under different handover edge strategies and show how to choose a better handover strategy to enhance its robustness.The further research work in this paper can also help design how to reduce the interdependence between systems,thereby further optimizing the interdependent network system’s structure to provide practical support for reducing the cascading failures.In the later work,we hope to explore our proposed strategies in the network model of real-world or close to real networks.

Edge Enhancement for Remote Sensing Image Using Norm Algorithm in L_2 Vector Space

The Householder transformation-norm structure function in L2 vector space of linear algebra is introduced, and the edge enhancement for remote sensing images is realized. The experiment result is compared with traditional Laplacian and Sobel edge enhancements and it shows that the effect of the new method is better than that of the traditional algorithms.

A new edge recognition technology based on the normalized vertical derivative of the total horizontal derivative for potential field data

边察觉和改进技术通常在用潜在的域数据认出地质的身体的边被使用。我们在场一种新边识别技术基于有边察觉和改进技术的功能的全部的水平衍生物的规范的垂直衍生物。首先，我们计算潜在地的数据的全部的水平衍生物(THDR ) 然后计算 n 顺序 THDR 的垂直衍生物(VDRn ) 。为 n 顺序垂直衍生物，全部的水平衍生物(PTHDR ) 的山峰价值用比 0 大的阀值价值被获得。这 PTHDR 能被用于边察觉。第二， PTHDR 价值被全部的水平衍生物划分并且由最大的价值使正常化。最后，我们使用了数字模型的不同类型验证新边识别技术的有效性和可靠性。

中兴通信增强型EDGE解决方案

在介绍增强型EDGE技术原则和关键技术的基础上,提出了中兴通讯增强型EDGE解决方案。

Performance Evaluation of Edge Detection Using Sobel, Homogeneity and Prewitt Algorithms

Edge detection considered as very important and fundamental tool in image processing. An image edge is a very sensitive place where the image information and details mostly placed on it. Different filters were used to detect and enhance these edges to improve the sharpness and raising the image clarity. The significance of this paper comes from the study, compare and evaluate the effects of three well-known edge detection techniques in a spatial domain, where this evaluation was performed using both subjective and objective manner to find out the best edge detection algorithm. The Sobel, Homogeneity and Prewitt algorithms were used on 2D gray-scale synthesis and real images in Jordan using C# programming language. According to the comparative results obtained using the three techniques, it was clearly found that Prewitt and Homogeneity algorithms performance were better than Sobel algorithm. Therefore, Prewitt and Homogeneity algorithms can be recommended as useful detection tools in edge detection.

The Test and Analysis of Edge Technology

Enhanced Data Rates for Global System for MobileCommunication (GSM) Evolution (EDGE) is a controversialtechnology for telecom operators. Basedon an actual network testing and analyzing, the paperdeals with the performance and actual impacts of thecompleted EDGE projects mainly through comparisonapproaches.

Contrast Enhancement of Radiographs Using Shift Invariant Wavelet Transform

A novel approach using shift invariant wavelet transform is presented for the contrast enhancement of radiographs. By exploiting cross-scale correlation among wavelet coefficients, edge information of radiographic images is extracted and protected, while noise is smoothed out in the wavelet domain. Radiographs are then reconstructed from the transform coefficients modified at multi-scales by nonlinear enhancement operator. The method can achieve effectively contrast enhancement and edge-preserved denoising simultaneously, yet it is capable of giving visually distinct images and offering considerable benefits in medical diagnosis.

A Hybrid Method for Edge Continuity Based on Pixel Neighbors Pattern Analysis (PNPA) for Remote Sensing Satellite Images

Edge enhancement is derived from lack of accurate result from edge detection techniques. The image which is captured from long distances carries a lot of noise and blur which causes edge discontinuity. Although some novel algorithms which are based on cellular neural network, fuzzy enhancement and binary morphology have shown accuracy in order to obtain refined edge but still the problem of edge discontinuity arises. Eliminating discontinuity of edge a hybrid technique is proposed based on pixel neighbors pattern analysis PNPA. In the technique Canny operator for initial edge detection, PNPA operation for edge enhancement are performed for remote sensing satellite image successively. The visual and subjective evaluation shows that the proposed PNPA operation can effectively eliminate the influence of edge discontinuity which occurred due to noise and blurr in original captured image, as comparing to existing edge segmenting processes.

Enhancement of Microcalcifications Based on Fractal Techniques in Mammograms

Microcalcification clusters in mammograms are an important early sign of breast cancer. The enhancement of microcalcifications in mammograms is one of the most important preprocessing techniques for the extraction of cluster microcalcifications. In this paper, we present a novel method for the enhancement of microcalcifications. Firstly, the initial microcalcification edges were extracted by using kirsch edge operator, and the discontinouse edges were linked by employing fractal technique. Then, the continuous closed edges of microcalcifications were filled by using seed filling algorithm. The pixel values of the filled region were replaced by the corresponding pixel values in the original image. Finally, the enhancement of microcalcifications in mammograms was achieved by adding the filled image to the original image. We evaluated the performance of our algorithm by using 50 regions of interesting (ROIs) with microcalcification clusters from DDSM database. The experiment results demonstrate that our CAD system can give better enhancement effect compared with other methods.

A New Method of CT Medical Images Contrast Enhancement

A new method of contrast enhancement is proposed in the paper using multiscale edge representation of images, and is applied to the field of CT medical image processing. Comparing to the traditional Window technique, our method is adaptive and meets the demand of radiology clinics more better. The clinical experiment results show the practicality and the potential applied value of our method in the field of CT medical images contrast enhancement.

MULTI-GRAY EDGE DETECTION BASED ON GENERALIZED FUZZY SET

In this letter, drawbacks of the classical algorithm to enhance the fuzzy contrast among adjacent regions are analyzed. Based on it, a new fuzzy enhancement algorithm and a linear fuzzy distribution that maps the gray images to corresponding generalized fuzzy set are proposed. Results of two examples illustrate that the algorithm is more effective and faster when used to detect the multi-level edges of images.

Distance Transform Based Enhancement for Linear Interpolated Images

An approach of distance map based image enhancement (DMIE) is proposed. It is applied to conventional interpolations to get sharp images. Edge detection is performed after images are interpolated by linear interpolations. To meet the two conditions set for DMIE, i. e., no abrupt changes and no overboosting, different boosting rate should be used in adjusting pixel intensities. When the boosting rate is determined by using the distance from enhanced pixels to nearest edges, edge-oriented image enhancement is obtained. By using Erosion technique, the range for pixel intensity adjustment is set. Over-enhancement is avoided by limiting the pixel intensities in enhancement within the range.A unified linear-time algorithm for distance transform is adopted to deal with the calculation of Euclidean distance of the images. Its computation complexity is O (N2 ). After the preparation, i. e.,distance transforming and erosion, the images get more and more sharpened while no over-boosting occurs by repeating the enhancement procedure. The simplicity of the enhancement operation makes DMIE suitable for enhancement rate adjusting.

面向病理图像分割的边缘感知网络

提出了一种针对病理切片图像的端到端语义分割方法--边缘感知网络(BPNet),以提高病理图像分割精度。BPNet网络首先在解码器阶段增加边缘感知模块,改善网络对于病理图像边缘的特征信息提取能力。然后,采用自适应通道注意力模块弥补不同层次特征间的语义差距,进一步加强网络的特征聚合能力。在此基础上,设计了一种基于结构和边缘的联合损失函数,以实现最佳的病理图像分割结果。在GlaS和MoNuSeg两个公开病理数据集上的分割实验结果表明,所提方法的Dice系数得分在两个数据集上分别达到92.21%和81.18%,有效提升了病理图像的分割精度。

基于前背景特性的逆光交通场景图像增强

由于逆光交通场景下采集到的图像存在前景亮度过低、背景亮度失真等特点,导致采集图像清晰度低、信息丢失严重和可识别性差,针对逆光图像前背景的不同特性,研究提出了一种分区域增强方法。对逆光图像使用最大类间方差法(OTSU法)分割图像的前景和背景;对逆光图像全局使用LIME方法,提升前景亮度的同时保持色彩的失真度;单独将背景部分RGB三个通道上的全局直方图均衡化结果一一映射到对应的限定区间内,提高了背景的对比度;使用Canny算子检测出前背景拼接处的黑边,根据黑边生成3个自适应的滤波模板对黑边进行分步均值滤波处理,消除了黑色边缘,提升了图像的视觉质量。在实验室自建的CHD_B数据集上,所提方法在4种常用客观评价指标上综合占优。实验结果表明,所提出的图像增强算法能有效地消除图像中的逆光现象。

增强彩色电子纸饱和度的误差扩散优化

为解决彩色电泳电子纸因粘滞阻力等引起的显示颜色饱和度低、边缘模糊等问题,本文提出基于HSL空间的彩色电子纸边缘增强误差扩散算法,以提高图像显示质量。该算法首先将去噪图像利用边缘检测算子得到边缘增强图像,结合边缘增强图像像素邻域平均灰度、像素与邻域灰度差异和像素邻域相似度得到新RGB图像像素值。接着,新RGB图像通过阈值处理过程得到16色阶RGB图像。最后,16色阶RGB图像转换到HSL空间,建立HSL和RGB色彩空间的转换模型,根据像素点的亮度和饱和度计算出调整因子,从而增强RGB图像饱和度。该算法与传统的误差扩散算法相比,信噪比PSNR提高了3.9%~26.7%,UCIQE提高了10.1%~48.2%,相似度SSIM提高了13.2%~25.4%。主观评价参考ITU-R BT.500-1标准设计实验计算Z得分,最终本文算法处理后图像在彩色电子纸上显示的图像细节和颜色更加接近原图,整体视觉效果更好。

融合对比度拉伸的地铁隧道环境图像复原算法

由于地铁隧道环境图像对比度低且细节模糊,在后续图像处理过程中难以从图像中准确提取关键的环境信息。为了解决该问题,提出一种融合对比度拉伸的模糊图像复原算法。将隧道图像转换到HSV颜色空间,根据V分量的灰度分布特点构建对比度拉伸模型对V分量进行增强,避免了传统低照度增强算法出现的过度增强问题,且能够自适应增强图像对比度。对增强后的V分量进行模糊类型分析,从而确定地铁隧道图像模糊类型,在此基础上将增强后的V分量图像划分为不同的区域,在各个区域分别选择符合条件的刀刃边缘线并对其扩散函数进行估计,同时以点扩散函数作为先验信息,采用非盲去卷积算法对各区域进行去模糊处理。融合H、S、V 3个分量,完成隧道环境低照度图像整体增强与复原。实验结果表明,所提算法能有效提高隧道图像整体及局部对比度,减少由高斯噪声引起的图像模糊,恢复图像中的细节信息。

一种低空遮挡场景感知增强方案研究

随着低空无人机业务的稳步快速增长,低空场景的通感一体化已成为低空经济发展中必不可少的关键一环,利用现有5G基站进行低空通感一体改造也成为研究热点。对于低空覆盖感知中的遮挡问题,提出利用一种电磁可控智能面及边缘提高感知性能,并提出了性能评估框架,以分析不同的感知方式和电磁操控机制的性能。仿真验证表明,在所研究的低空场景通感一体感知任务中,电磁操控机制的加入可以有效提升无人机感知概率,同时被动感知方式优于主动感知方式,而典型空-地场景下所提出的绕射增强结构的性能表现也优于智能反射面。

基于特征融合的低剂量CT图像降噪方法

近年来低剂量CT(Low Dose CT,LDCT)被广泛应用于临床诊断中,但LDCT会产生不规则的噪声。已有的降噪方法往往缺乏对全局特征信息的考虑,以及不注重边缘特征信息和重建图像的视觉效果。为此,提出了一种基于特征融合的低剂量CT图像降噪方法。首先,利用Transformer优异的全局感受野提取图像的全局特征信息,并利用卷积神经网络(Convolutional Neural Network,CNN)良好的局部特征提取能力提取图像的局部特征信息。在Transformer模块中加入维度变换思想,以更好地抑制噪声;在CNN模块中使用稠密连接的方式将浅层网络的特征信息复用于深层网络中,以此保存更多的特征信息。其次,为了获取更加丰富的图像细节特征,使用了改进的索伯边缘增强算子来加强模型对边缘特征信息的提取能力。最后,将Transformer模块和CNN模块获取的特征信息进行融合并输出重建图像。此外,为了使降噪重建后的图像有更好的质量和视觉效果,设计了一个多尺度复合损失函数。实验表明:在AAPM-Mayo数据集的降噪实验中,与当前主流的LDCT图像降噪方法相比,本文方法取得了更好的降噪效果。

基于CNN跨层融合结构的边缘检测算法

传统边缘检测算法难以处理复杂的图像,而现有基于深度的边缘检测模型,其检测结果往往存在边缘定位错误和信息丢失等现象.针对此类问题,提出一种基于RCF的高精度的边缘检测算法RCF-CLF.首先,引入HDC结构设计用于避免因叠加相同膨胀卷积而引起的网格效应;其次,设计了一种特征增强结构,旨在融合多尺度信息、扩大感受野;然后,设计了跨层融合结构,将高层信息和低层信息融合,用于提取准确的边缘信息;最后,引入注意力机制CBAM,通过聚焦物体边缘区域,抑制非边缘区域,从而提高网络对边缘信息的提取能力.本文在BSDS500和BIPED数据集上评估所提出的方法,与RCF算法相比,在BIPED数据集上,主要指标ODS、OIS和AP分别达到了0.893、0.901和0.945,提高了近5个百分点,在BSDS500数据集上,主要指标也有所提升.此外,与其他同类算法相比,本文算法也具有一定的优势,可以实现更加准确的边缘定位.