Improved visual/infrared colour fusion method with double‐opponency colour constancy mechanism

IET Image Processing, Journal Year: 2018, Volume and Issue: 12(9), P. 1560 - 1566

Published: April 11, 2018

To solve problems that colour distortion and low resolution of infrared visible fusion images, the authors propose a method based on double‐opponency constancy mechanism human vision. First, Waxman's which imitated neuro‐dynamics rattlesnake bimodal cell is used to fuse light with source multi‐band images generate pseudo‐colour images. Second, double‐opponent computation model Rodieck's double difference‐of‐Gaussian proposed obtain estimate an illuminant Finally, are corrected by diagonal transformation cone adaptive mechanism. They also use three‐dimensional RGB histogram analyse distribution In comparison experiments other approaches using histogram, one can see gives image coinciding well natural satisfies perception needs.

Language: Английский

基于拮抗特性模型的夜视微光图像与红外图像彩色融合

Réngōng zhìnéng qiányán yǔ yìngyòng, Journal Year: 2024, Volume and Issue: 1(1), P. 45 - 53

Published: May 28, 2024

融合可以有效地利用可见光图像的色彩信息得到较好的可视效果，又可以充分利用红外图像获得人眼无法观察到的红外信息，具有广阔的应用前景。论文首先采用拮抗特性模型中的中心-周边对抗网络对红外与微光图像进行增强，利用区域生长方法对增强后红外图像进行分割，根据分割后各区域亮度得到目标图像。在融合阶段加入两种融合运算：采用选择运算将目标图像信息融合到增强后微光背景中作为亮度通道的输入；利用拮抗特性模型将经采用区域生长法得到红外图像和增强后的微光图像进行融合，作为饱和度通道的输入。同时，直接将增强的微光图像送入调色通道, 作为彩色融合图像的背景进行调色，最后通过彩色重映射并加以显示。实验结果获得的图像具有较好的目标指示特性，色彩更适合人眼观察，有利于提高对目标情景的感知能力。