隔点差分自适应阈值红外图像分割算法的FPGA实现  

Interval Dot Difference with Adaptive Threshold Segmentation Algorithm Using FPGA in IR

作　　者：朱军[1] 史彩成[1] 何佩琨[1] 

机构地区：[1]北京理工大学电子工程系,北京100081

出　　处：《激光与红外》2005年第2期134-136,共3页Laser & Infrared

基　　金：国家 863计划项目(No. 2002AA813032)。

摘　　要：采用模块化设计和利用现场可编程逻辑 (FPGA)并行流水处理特点,设计并实现了隔点差分自适应阈值红外图像分割算法,算法由隔点差分算法和自适应阈值两个模块组成。隔点差分算法模块根据同步与时钟流水计算隔点梯度,用自适应阈值进行图像分割; 自适应阈值模块根据背景的统计规律和图像背景的变化实时调整分割阈值。仿真和实验表明:采用FPGA实现此算法的处理延时短,满足实时处理的要求。With the modularization method and utilizes characteristic of parallelity and pipeline of FPGA, interval dot difference with adaptive threshold segmentation algorithm is designed and implemented. The implementation consists of two modules: interval difference algorithm module and adaptive threshold algorithm module. The former carries out pipeline computation of interposing gradient with synchronizing time sequence and clock; the latter conducts adaptive adjustment of segmentation thresholds in real time with the change of image background according to statistic characteristic of background. Results of simulation and experiments show that this design can process image segmentation with short delay and satisfy the real-time processing requirement.

关 键 词：图像分割 隔点差分自适应阈值 现场可编程逻辑 

分 类 号：TN957[电子电信—信号与信息处理]