多重光照下彩色数字图像色差在线检测仿真  

Multi-Light Multi-Color Digital Image Color Difference Online Detection Simulation

作　　者：钟小莉[1] 樊吉亮[1] 李永宁[1] ZHONG Xiao-li;FAN Ji-liang;LI Yong-ning(Computer School, Qinghai Nationalities University, Xining Qinghai 810007, China)

机构地区：[1]青海民族大学计算机学院

出　　处：《计算机仿真》2019年第8期454-458,共5页Computer Simulation

基　　金：基于立体图像智能分割技术应用的算法研究(2018XJY01)

摘　　要：对多重光照下的彩色数字图像色差进行在线检测,在提高图像利用率方面就要重要意义。在对彩色数字图像色差进行检测时,需对图像反射和照明的变化率进行获取,确定检测所需隶属函数。传统方法主要根据图像色调及饱和度进行检测,忽略了隶属函数的影响,导致检测精度低、效率差的问题。现提出基于模糊C均值聚类的图像色差在线检测方法。利用彩色数字图像的反射分量模型和照明分量模型,计算光照彩色数字图像反射和照明时色差的变化率,在彩色数字图像亮度直方图中建立图像映射函数,利用映射函数完成对彩色数字图像的校正。根据校正结果,结合彩色数字图像的色调、饱和度和明度对彩色数字图像颜色进行分割,并将分割结果代入彩色数字图像聚类所需隶属函数中,引入模糊C均值聚类,实现对彩色数字图像色差的在线检测。仿真结果表明,提出方法提高了多重光照下彩色数字图像色差在线检测的精度和效率。The online detection for the chromatic aberration of color digital image under multi illumination is very important in improving image utilization. When detecting the chromatic aberration of color digital images, it is necessary to obtain the change rate of image reflection and illumination, so that the membership functions required for detection can be determined. The traditional method mainly detects the chromatic aberration based on the hue and saturation level of image, but ignores the influence of membership function, leading to the low detection precision and poor efficiency. Therefore, this paper puts forward a method to detect the color difference on line based on fuzzy C means clustering. At first, the reflection component model and the illumination component model of color digital image was used to calculate the change rate of color difference when the light color digital image was reflected and illuminated. Then, the image mapping function was established in the brightness histogram of color digital image. The mapping function was used to correct the color digital image. According to the correction results, the color of digital image was segmented by combining the hue, saturation and brightness of color digital image, and then the segmentation results were plugged in the membership function needed by the color digital image clustering. Finally, fuzzy C means clustering was introduced. Thus, the on-line detection of chromatic aberration of color digital images was achieved. Simulation results show that the proposed method improves the accuracy and efficiency of online detection for color difference of color digital image under multi illumination.

关 键 词：多重光照下 彩色数字图像 色差 在线检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]