...
首页> 外文期刊>International Journal of Engineering Science and Technology >A SURVEY ON FPGA PROTOTYPING OF DIGITAL ARCHITECTURES OF EDGE DETECTION TECHNIQUES
【24h】

A SURVEY ON FPGA PROTOTYPING OF DIGITAL ARCHITECTURES OF EDGE DETECTION TECHNIQUES

机译:边缘检测技术数字体系结构的FPGA原型研究

获取原文
           

摘要

In image processing, edge detection and extraction for noisy as well as filtered images plays a vital role. This work presents a broad literature survey of FPGA Prototyping for feature extraction of different kind of images and videos focusing on edge detection techniques and their evolution with respect to improvement in architectural features. Further, the work reported so far has been compared for system performance with a consideration on multiple factors such as frequency, time, power, latency, speed, throughput, cost, complexity and hardware utilization of various families of FPGAs. The methodology opted for detecting edges and image/video specifications also influence the system performance. Sobel, Canny, Prewitt, Robert are the most widely used techniques for enhancing edges which have been modified and implemented on various FPGAs to achieve better performance parameters. Distributed Canny Edge detection, a block based technique and Multi-resolution Self-adapting architecture have reported reduced latency, best timing performance, increased throughput and enhanced edge detection for noisy images. Besides these some techniques like Residue number system assisted edge detection, Neuro Fuzzy system based detection, Spiking Neural Networks, LEDges, Beamlet Transform Algorithm, Clifford algebra have also proved to achieve better performance in some medical imaging and object tracking applications.
机译:在图像处理中,噪声和滤波图像的边缘检测和提取起着至关重要的作用。这项工作提供了针对FPGA原型的广泛文献调查,以针对不同类型的图像和视频进行特征提取,重点是边缘检测技术及其在改善体系结构功能方面的发展。此外,已经比较了迄今为止报告的工作的系统性能,并考虑了多种因素,例如频率,时间,功率,等待时间,速度,吞吐量,成本,复杂性和各种FPGA系列的硬件利用率。选择用于检测边缘和图像/视频规格的方法也会影响系统性能。 Sobel,Canny,Prewitt,Robert是增强边缘的最广泛使用的技术,这些技术已在各种FPGA上进行了修改和实现,以实现更好的性能参数。分布式Canny边缘检测,一种基于块的技术和多分辨率自适应体系结构已报告了减少的延迟,最佳的定时性能,增加的吞吐量以及增强的用于噪点图像的边缘检测。除了这些技术外,诸如残数系统辅助边缘检测,基于神经模糊系统的检测,尖峰神经网络,LEDges,Beamlet变换算法,Clifford代数等还被证明在某些医学成像和对象跟踪应用中具有更好的性能。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号