基于改进灰狼优化的红外与可见光图像融合

Infrared and Visible Image Fusion Based on Improved Gray Wolf Optimization

将图像融合运用于检测与跟踪领域需要融合图像显示清晰的目标,传统的优化类融合算法存在目标信息不完整的问题,对此本文提出一种基于改进灰狼优化(Gray Wolf Optimization,GWO)结合边缘特征的图像融合方法。将图像分解为细节层与粗糙层后,对细节图像使用优化权重进行融合,再融合细节层与粗糙层,最后执行对比度有限自适应直方图均衡增强融合图像。其中优化权重通过改进的灰狼优化获得,通过融合边缘信息获得权重取值范围,并且对灰狼优化引入交叉操作改进优化效果。实验对比图像全局与目标局部的标准差、信息熵、平均梯度、空间频率,本文方法的性能在目标局部熵、标准差上大大优于其他方法,在全局指标上也有很好的表现。

When image fusion is applied to the detection and tracking field, it need to show clear objectives. The traditional optimization fusion algorithm has the problem of incomplete target information. This study proposes an image fusion method based on improved gray wolf optimization(GWO) combined with edge features. After the image is decomposed into the detail layer and coarse layer, the detail image is fused with the optimized weight, and the detail layer and coarse layer are fused. Finally, the contrast limited adaptive histogram equalization enhancement fusion image is performed. The optimization weight is obtained through the improved GWO. The weight value range is obtained by blending the edge information, and the crossover operation is introduced to improve the optimization effect of GWO. The global and target local standard deviation, information entropy, average gradient, spatial frequency are experimentally compared for different methods. The results show that the proposed method performed considerably better than other methods in terms of target local entropy and standard deviation, and also had good performance in terms of global indicators.