Application of UAV image homogenization based on improved Retinex algorithm
The traditional single-scale Retinex homogenization algorithm uses Gaussi Low-pass Filter.When the Gaussian filter function is applied to the spectrum of the UAV image,the high-frequency infor-mation will be truncated,some image details will be lost,which will affect the light and color homogeni-zation effect.So a Retinex homogenization algorithm based on Butterworth Low-pass Filter is proposed in this paper.Butterworth Low-pass Filter is used to extract illumination component map,which can protect image pixels and details and improve the effect of homogenization of light and color.The test was carried out using one scene UAV images to verify the reliability of the algorithm,and it is compared with other four methods of light and color homogenization.It is concluded that the algorithm proposed in this paper has obvious advantages in visual effect and quantitative evaluation index.The texture and details of the image can be maintained while the effect of light and color homogenization is improved.
UAV imagehomogenization of light and colorRetinexButterworth Low-pass Filtering
NETL
NSTL
万方数据
