摘要
介绍了基于嵌入式Linux系统的S3C2410平台,并在该平台上构建了实现运动图像自动监控的软件开发环境。通过驱动COMS图像传感器,利用Video4Linux在该平台上实现了运动图像数据的捕获。在运动图像识别方面,对帧差法进行了分析探讨并作了改进,使得该算法更易于在ARM平台上实现。通过引入灰度直方图判断有无运动目标出现,利用阈值对图像二值化处理,分割效果良好,实验结果表明该算法适于应用在嵌入式实时视频监控场合。
The S3C2410 platform based on the embedded Linux is introduced in this paper. And the software environment is built on the platform which can monitor moving images automatically. Then it expatiates how to drive COMS image sensor and how to achieve image gathering by Video4Linux. In terms of moving images identifying, Frame-Difference method is analyzed, discussed and improved to make it easier running in ARM platform. It judges whether there is a moving object by grayhistogram and images are two-numeral processed by threshold and the effect of the segmentation is good. The validity of the method is confirmed by experimental results applying to real-time embedded video monitoring.
出处
《电子测试》
2009年第2期54-58,86,共6页
Electronic Test