基于双口RAM的新型FPGA质心提取算法设计  

A New FPGA Algorithm of Centroid Extraction Based on Dual-port RAM

作　　者：马超[1,2] 章明朝[1] 李佩玥[1] 周跃[1] 隋永新[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所应用光学国家重点实验室,长春130033 [2]中国科学院大学大珩学院,北京100039

出　　处：《计算机测量与控制》2015年第4期1289-1291,1295,共4页Computer Measurement &Control

基　　金：国家高技术研究发展计划(863计划)(2012AA0XXX07)

摘　　要：针对大多数现有FPGA质心提取算法不适用于大量连通区图像的问题,设计了一套基于双口RAM的新型FPGA质心提取算法;在保持普通FPGA质心算法无需存储整幅图像和只扫描一次图像便可得到质心坐标的特点的前提下,巧妙地使用3个双口RAM作为算法的存储介质,避免了资源的大量消耗,由于没有了资源上的限制,新型算法不再局限于少量连通区图像,已然能够对千个以上的连通区图像进行处理;最后进行了一系列的实验分析:与Matlab算法的处理结果对比,验证了设计的正确性;与普通FPGA质心算法的实验对比,说明新型算法在节省大量资源的同时还使得算法效率提高了一倍。In order to solve the problem that most existing FPGA centroid extraction algorithm does not apply to images with large number of connected regions, a new FPGA centroid extraction algorithm based on dual port RAM was designed. In the ease of keeping ordinary FPGA centroid algorithm＇ s good features like that the coordinates of centroid could be obtained by only scanned the image once and with no need to store the whole image, algorithm use 3 dual-port RAM to substitute the ordinary storage media, that reduce the serious consumption of resources, and overcome the shortcomings of the ordinary algorithm, now the new algorithm could be used to the images with even thousands of connected regions. At the last, a series of experimental and comparative analysis was done to prove the correctness of the design. And compared with the ordinary algorithm, the new algorithm saves a great number of logic resources while doubled the speed of operation.

关 键 词：质心提取 双口随机存取存储器 连通区 像素标记 现场可编程门阵列 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]