Research On Fusion Method Of Carbonized Wood Micro Image Based On Super Resolution Multi-Focus

Studying carbonized wood from a microscopic perspective has certain scientific research significance.Analyzing and detecting carbonized wood from a microscopic perspective can observe the corresponding structure of the cells on the surface of the carbonized wooden and the interaction between the cells.For further discussion on the carbonized wood The cause of surface cracking and shrinkage has a great auxiliary effect.Obtaining clear image data of carbonized wood surface is the first condition for the research of carbonized wood from the micro level.The complicated carbonization experiment environment and the insufficiency of general micro-camera instruments make it difficult to obtain clear image data information of carbonized wood surface.To solve this problem,this paper proposes a multi-focus image fusion algorithm based on super-resolution.This article does the following work:(1)According to the requirements of the carbonization experiment environment,a wood carbonization system that can simulate the real wood carbonization environment was built.Controlling the two environmental parameters of carbonization temperature and carbonization time in the carbonization system by means of control variables,a large number of wood carbonization experiments were carried out,and a large amount of microscopic carbonized wood image data was obtained by using a digital microscope in the carbonization system.(2)Introduce the basic contents of the three image fusion algorithms of super-resolution method,stationary wavelet transform method and discrete wavelet transform method,and study and compare the advantages,disadvantages,application scope and basic characteristics of these three algorithms.The image fusion simulation is carried out for these three image fusion algorithms;five evaluation standards for fusion image quality are proposed:root mean square error,peak signal-to-noise ratio,average gradient,petrovics index,and mutual information.And the simulation results of the three image fusion algorithms are evaluated,and the specific characteristics of the three image fusion algorithms are analyzed according to the evaluation results.(3)Combining the advantages of the three image fusion algorithms mentioned in this article,a super-resolution multi-focus image fusion algorithm is proposed.This algorithm first performs a single-frame super-resolution processing of all three source images with bi-cubic interpolation.Enhance details such as source image contrast.Then,a stationary wavelet transform is implemented on these high-resolution source images,and the source image is divided into four subbands.Aiming at the problems of chaotic detail information and redundant structure information of the source images contained in these sub-bands,principal component analysis is used to select the maximum signal-to-noise ratio of each sub-band of the source image for image fusion.Finally,the inverse stationary wavelet transform is used to reconstruct the fusion sub-band to obtain a high-quality fusion image.Using this algorithm,the collected micro-images of carbonized wood were fused to obtain clearer image data,and the quality evaluation indexes of the fused images were relatively good.