Testing and Simulation of Fuzzy Detail Self adjustment Enhancement in Weak Illumination Images
Due to the strong randomness of natural light,the images captured by cameras under weak light are of-ten blurry,and it is also difficult to obtain detail information.To address this problem,this paper presented a self-reg-ulating enhancement method for blurred details in weak lit images.At first,a similarity threshold function was intro-duced into the non-local equalization filtering algorithm to remove noise from blurred images.Then,the fractional or-der total variation model and the non-local total variation model were used to construct a deblurring model.After the image deblurring,the preprocessed images were input into a multi-scale convolutional neural network model.Finally,the image features were extracted and fused through global and local paths,thereby achieving self-regulating enhance-ment of blurred details in weak-illumination images.The experimental results show that the proposed method has smaller mean square error,higher structural similarity and peak signal-to-noise ratio in synthetic image enhancement,while in real image enhancement,it has higher evaluation metric of natural image quality and higher contrast gain.Meanwhile,the entropy is closer to 8,and the image processing time is only 0.745ms.
Weak light imageBlur detailsSelf-regulation enhancementNon local mean filtering NLMFCon-volutional neural network
NETL
NSTL
万方数据
