基于颜色梯度融合高斯模型的光照突变干扰消除

ELIMINATING LIGHT MUTATION INTERFERENCE BASED ON FUSION OF COLOUR GRADIENT AND GAUSSIAN MODEL

光照突变对图像处理中的图像检测与识别具有重要影响。在分析光照干扰的突变特性基础上,提出一种基于颜色梯度融合高斯模型的光照突变干扰消除方法。首先通过在传统的高斯图像模型中加入颜色突变梯度融合算子,对传统的模型进行合理的优化,尝试性地消除光照突变对模型稳定性的影响;然后利用背景梯度分布获取的梯度矢量概率对背景区域进行合理的区分。后期的计算机仿真实验结果表明:优化后的模型对于消除图像在光照突变环境下的干扰效果较好,实验数据也支持了这一观点。

Light mutation has significant impact on image detection and recognition in image processing. Based on analysing the mutation characteristics of light interferences, we put forward a method for eliminating the light mutation interference which is based on fusion of colour gradient and Gaussian model. First, we optimise the traditional model reasonably through putting the colour mutation gradient fusion operators in traditional Gaussian image model, and try to eliminate the impact of light mutation on the stability of the model ; then we use gradient vector probability obtained with the distribution of background gradient to reasonably differentiate the background region. The result of computer simulation experiment in later stage demonstrates that the optimised model has good effect in eliminating the image interference in environment of illumination mutation, the experiment data also supports this opinion.