基于Retinex理论的船舶图像照度提取算法

A ship image illumination extraction algorithm based on Retinex theory

针对船舶环境亮度的不同,通过摄像机获取船舱环境图片信息,将HIS模型与Retinex理论结合获取图像照度模型,从而自适应控制船舱LED灯照度。将RGB空间中的彩色图像映射到HIS空间,根据瑞利散射特性获取不同颜色通道权重,得到改进后HIS亮度分量模型;应用Retinex理论将亮度分量与高斯中心环绕函数做卷积运算,最终得到图像照度估计。实验结果表明,与传统算法相比,本文算法减少了颜色分量干扰,能获取到更为精确的照度信息。

According to the different brightness of the vessel environment,environment image information is obtained by the camera,and the HIS model is combined with the Retinex theory to obtain the image illumination model,thereby adaptively controlling the illumination of the LED light of the cabin.Firstly,the weights of different color channels are obtained by the iso-illuminance curve method,and the RGB color image is converted into the HIS space to obtain the improved luminance component model.Then,the Retinex theory is used to convolve the luminance component with the Gaussian center surround function,and finally the image illumination is obtained.The experimental results show that compared with the traditional algorithm,this algorithm can reduce the interference of color component and obtain more accurate illumination information.