紫外序列图像中目标的提取

Target extraction from the ultraviolet image sequences

分析了"日盲"紫外ICCD(增强型电荷耦合装置)探测系统所采集紫外图像中噪声和目标的特点,基于紫外图像的特点,提出了一种紫外序列图像中目标提取的方法。该方法首先采用时域递归低通滤波算法对紫外图像进行降噪处理,有效抑制了图像中的随机噪声,提高了图像的对比度。然后运用自适应阈值分割算法对目标进行了提取。实验结果表明,该方法能较好地检测出紫外序列图像中的目标,具有较强的噪声抑制能力。

The noise and target characteristics of the ultraviolet image detected by the Ultraviolet Intensifier Charge Coupled Device （UV ICCD） were analyzed. Based on the ultraviolet image characteristic and the noise speciality, a method for target extraction of the solar blind ultraviolet image sequences was presented in this paper. At first, a time-recursive low-pass filtering algorithm was adopted to decrease the noise of the ultraviolet image. The random noise of the ultraviolet image sequences was suppressed. At the same time, the contrast of the ultraviolet image was improved. Then the targets were extracted by the algorithm of adaptive threshold segmentation. Experiment results demonstrate that this method is effective to detect the target of the ultraviolet image sequences, and it has strong ability for restraining the noise.