基于Hough变换的红外图像的目标分割

Object Segmentation of Infrared Image Based on Hough Transform

红外图像是根据目标与背景的温度差异而形成,不受光照等其他条件的影响,可全天工作,被广泛应用在各个领域中。但是由于目标和背景的热辐射所造产生的红外图像的分辨率低,对比度差等特点,使红外图像的分割变得较为困难。Hough变换具有对图像中直线的残缺部分、噪声以及其他共存的非直线结构不敏感的特性,具有较强的抗干扰性和鲁棒性。本文提出一种基于Hough变换的红外图像规则目标的检测方法。通过观察红外图像中目标分布情况,估计出目标边界的曲线方程,对该方程进行Hough变换,从而检测出目标并进行分割。利用该方法避免了由于红外图像的目标与背景对比度差、边缘模糊、视觉效果较差等原因造成的不易检测边缘曲线的问题。经实验证明,基于Hough变换的规则的红外目标的检测方法具有可行性,且检测效果好。

Infrared image is formed according to the temperature difference between the target and the background. It is not affected by other conditions,such as light and it can work all day. So it is widely used in various fields. However,due to the low resolution and low contrast of infrared image generated by the thermal radiation of the target and background,the segmentation of infrared image becomes more difficult. Hough transform has a strong anti- interference and robustness,which is not sensitive to the incomplete part of the image,noise and other non- linear structures. In this paper,a method for the detection of regular targets in infrared image based on Hough transform is proposed. By observing the distribution of target in infrared image,the curve equation of the target boundary is estimated,then the Hough transform is carried out to detect the target and segment the target. By using this method,it is avoided that the target of infrared image is not easy to detect the edge curve due to the difference of background contrast,blurred edge and poor visual effect. The experimental results show that the infrared target detection method based on Hough transform is feasible,and the detection effect is good.