红外热像校正技术的仿真研究及其实时应用——双阈值分割算法结合伪彩色变换

Simulation and Real-time Application of Infrared Thermal Image Rectification Technology of Rectification Algorithm of Double Threshold Segmentation in Pseudo-color Conversion

近年来,红外热像可应用于变压器等电力设备的监测和过热点定位,受到现代电力行业的关注,但红外热像仪呈现的温度分布多为灰度图,不利于人眼观察和辨识,而且图像模糊,影响检测质量,因此,实现红外热图像彩色显示和校正是必要的。文章提出了一种图像增强的方法,即双阈值分割算法,经过算法的仿真实验后将其应用在红外热像仪采集的变压器灰度图的伪彩色变换中,并在DSP642+FPGA组成的处理系统运行了这些算法。结果表明:伪彩色算法结合图像校正算法可较为精确地实现色彩变换,获得较好的视觉效果,可比原图显示更多的图像细节,解决了红外图像处理中细节不清楚的问题。

Recently, the technology of infrared thermal imaging is being paid more attention tofor its advantage of monitoring the image and localizing the faults of electrical equipment such as transformer. However, the temperature picture supplied by the infrared imager is gray and blurring, not easy to eye observation and monitoring, so it is necessary for the infrared thermal image to be displayed in a pseudo-color manner and to be rectified.Here, an algorithm of image enhancement termed as double threshold segmentation was proposed, and was applied to the pseudo-color conversion for original gray images of electrical transformer after algorithm simulation. Additionally, the algorithm ran in the double-chip processing system of DSP642＋FPGA. It was resulted that the combining algorithms of pseudo-color conversion and rectification could complete precisely the color conversion for a better visual color image to be shown with more details than the original gray image, and the problem of unclear image details was solved in infrared image processing.