摘要
针对现有自适应中值滤波器运行耗时长以及在高椒盐噪声密度下去噪性能低的问题,提出一种基于FPGA的自适应中值滤波器优化方法,对铁轨图像进行滤波研究。加入像素噪声阈值判断和二次求中值的方法,提高图像在大椒盐噪声密度下的处理能力,在大窗口中只对外围像素求中值减少迭代处理,降低算法的复杂度和运算量,利用FPGA的流水线结构和并行处理能力,加速算法的运行。与现有仿真滤波效果最优的中值滤波器相比,优化方法图像细节保持更加完整,运算量降低15%,峰值信噪比值平均提高0.9 dB,采用FPGA运行优化算法将耗时降低到了毫秒级,为铁轨图像实时处理应用提供了基础。
An optimization method of adaptive median filter based on FPGA is proposed to filter the railway image on the problem of adaptive median filter with long time-consuming operation and poor denoising performance at high salt and pepper noise density.The method of pixel noise threshold judgment and secondary median calculation is added to improve the image processing ability under the large salt and pepper noise density.The median value of the peripheral pixels is found in the large window to reduce iterative processing,which could reduce the complexity and calculation amount of the algorithm.FPGA pipeline structure and parallel processing capabilities are used to speed up the operation of the algorithm.Compared with the existing median filter with the best simulation filtering effect,the optimization method maintains more complete image details,reduces the amount of calculation by 20%,and increases the peak signal-to-noise ratio by an average of 0.9 dB.Using FPGA to run optimization algorithms reduces the time consumption to the millisecond level,which provides a basis for the real-time processing of railway image.
作者
王丽红
胡长宏
范鲜红
高春歌
张晓峻
孙晶华
WANG Li-hong;HU Chang-hong;FAN Xian-hong;GAO Chun-ge;ZHANG Xiao-jun;SUN Jing-hua(College of Physics,Jilin University,Changchun 130012,China;School of Physics and Optoelectronic Engineering,Harbin Engineering University,Harbin 15001,China)
出处
《哈尔滨理工大学学报》
CAS
北大核心
2021年第5期68-75,共8页
Journal of Harbin University of Science and Technology
基金
国家自然科学基金(61775044)
中央高校基本科研专项资金项目(3072019CF2515).
