Adaptive Infrared Thermal Image Enhancement Based on Retinex

Aiming at the features of infrared image,a modified algorithm based on Retinex is proposed.A nonlinear transformation function is introduced to amend illumination and reflection components of infrared image.The global contrast enhancement function helps stretch illumination component which leads to improvement of the whole visual effect.Meanwhile the nonlinear sigmoid function is used to gray stretching locally so that better contrast of the image is achieved.Qualitative analysis and quantitative analysis are conducted by comparing the proposed algorithm and some classic Retinexalgorithms.It can be concluded from the comparison that proposed algorithm performs better than classic Retinex algorithms on enhancingtheinfrared image.the proposed algorithm can enhance contrast of infrared images and highlight the edge and details of the image effectively.