| 基于细胞免疫的红外图像分割与FPGA实现 |
| 投稿时间：2020-03-27 修订日期：2020-04-11 点此下载全文 |
| 引用本文：李大华,王宇,高强,于晓.基于细胞免疫的红外图像分割与FPGA实现[J].红外,2020,41(6):41~48 |
| 摘要点击次数: 60 |
| 全文下载次数: 44 |
|
|
| 中文摘要:在工业生产中，高温蒸汽管道通常被用于输送高温蒸汽、高温工业废水等，但为了安全，管道通常被放置在环境比较复杂的地方，不利于工人对管道的检测与维护。快速地定位复杂背景下蒸汽管道的位置并对周围环境进行区分，已经成了一个亟待解决的问题。由于最大类间方差（Otsu）算法不能满足上述要求，基于细胞免疫机制提出了一种改进的Otsu算法，该算法通过红外图像中管道以及复杂背景的特征，能够计算出两个不同的阈值并将其分别用于图像中管道的提取与复杂背景的区分。借助QuartusⅡ软件搭建了基于FPGA的软硬件系统平台，实现了数据通信传输，并对改进的Otsu算法进行验证。实验结果表明，该算法应用在红外管道图像中能取得较好的效果。与几种边缘检测算子和经典Otsu算法相比，无论是在管道的分割，还是复杂背景的区分，本文算法都具有较高的真阳率(True Positive Rate, TPR)和较低的假阳率(False Positive Rate, FPR)。 |
| 中文关键词:红外图像 FPGA 细胞免疫 阈值分割 |
| |
| Infrared Image Segmentation Based on Cellular Immunity and FPGA Implementation |
|
|
| Abstract:In industrial production, high-temperature steam pipeline is usually used to transport high-temperature steam, high-temperature industrial wastewater, etc, but for the sake of safety, pipelines are usually placed in places with complex environment, which is not conducive to the detection and maintenance of the pipeline by workers. Quickly locate the location of steam pipeline in complex background, and divide the surrounding environment into a problem to be solved. Because the Otsu algorithm can not meet the above requirements, an improved Otsu algorithm is proposed based on the cellular immune mechanism. The algorithm can calculate two different thresholds through the characteristics of the pipeline and the complex background in the infrared image, which can be used to distinguish the pipeline extraction and the complex background in the image. In this paper, with the help of Quartus II software, a FPGA based software and hardware system platform is built to realize data communication and transmission, and the improved Otsu algorithm is verified. The experimental results show that the algorithm can achieve better results in infrared pipeline image. Compared with several edge detection operators and classical Otsu algorithm, this algorithm has higher true positive rate (TPR) and lower false positive rate (FPR), no matter in pipeline segmentation or in complex background. |
| keywords:infrared image FPGA cellular immunity threshold segmentation |
| 查看全文 HTML 查看/发表评论 下载PDF阅读器 |
|
|
|
|
|