论文部分内容阅读
提出能同时消除亮度和反差分布不均匀现象的分块均值方差变换法,重点对插值匀光法、MASK匀光法、同态滤波匀光法、Retinex匀光法以及分块均值方差变换法的校正效果进行了实验比较。研究表明:分块均值方差变换法既能消除亮度分布不均匀现象,又能消除反差分布不均匀现象,而MASK匀光法、插值匀光法、Retinex匀光法只能解决亮度分布不均匀问题,同态滤波匀光法仅能减弱亮度分布不均匀现象;在匀光效果方面,分块均值方差变换法最好,MASK匀光法次之,插值匀光法紧随其后,接下来是Retinex匀光法,同态滤波匀光法最差。
This paper proposes a block-wise mean-variance transformation method that can eliminate the unevenness of luminance and contrast at the same time. Emphasis is put on the methods of interpolation uniform light method, MASK uniform light method, homomorphic filter uniform light method, Retinex uniform light method and block mean variance transform method Correction effect of the experimental comparison. The research shows that the block mean variance transform method can not only eliminate the unevenness of luminance distribution, but also eliminate the uneven distribution of contrast. However, the MASK uniform method, the interpolation uniform method and the Retinex uniform method can only solve the problem of uneven luminance distribution , The homomorphic filter uniform light method can only weaken the unevenness of the luminance distribution; in the uniform light effect, the block mean variance transformation method is the best, followed by the MASK uniform light method followed by the interpolated uniform light method, followed by Retinex uniform method, homomorphic filter method even the worst.