Multispectral Image Registration And Fast Filtering Methods

Multispectral images contain spectral data of a series of specific electromagnetic frequency bands.These images are rich in spectral information that can accurately characterize the surface color of an object.Multispectral images can be acquired by using multispectral imaging technique,which divides the spectrum into multiple spectral bands and takes sample of each band.Multispectral imaging technique is usually applied in textile printing and dyeing industry,where textile color measurement and evaluation of precision are strictly required to obtain accurate color information.Multispectral Imaging Color Measurement(ICM)system is a multispectral imaging system based on filtering method.The system mainly consists of light sources,a monochromatic camera and a filter-wheel with multiple optical band-pass filters.The ICM system first captures the camera response of each channel and then uses reflectance reconstruction algorithm to achieve a spectral measurement in camera resolution level.By this means the system accurately acquires the surface color of the object.The system provides a color management solution for the textile industry,including multispectral image measurement,color difference analysis,color quality control and so on.In this paper,a new method of image registration based on affine transform and Levenberg-Marquardt(LM)algorithm is proposed to solve the problem of color difference analysis for textile fabric images captured by the ICM system.In a registration perspective,the standard image and the batch image are first matched by the proposed method and then taken for analysis of spatial color difference.The deformation of the textile image captured by ICM system,including translation,rotation,scaling and shearing,conforms to the classic model of affine transform.The proposed method first calculates the integral curves of the log-polar magnitude spectra of the two images for mapping.The original problem of solving the transform matrix is then converted to a numerical problem of nonlinear least squares fittingwhere the LM algorithm is employed for optimal value searching.Besides,block registration is introduced to achieve more accurate registration.The experimental results show better registration accuracy of the proposed algorithm compared with the traditional Fourier-Mellin algorithm as well as the scale invariant feature transform(SIFT)based feature point matching method.It also solves the registration problem for textile images with periodic patterns effectively,which contributes to the process of color difference evaluation.In addition,fast filtering algorithms are studied for engineering applications of the ICM system.The image processing module belongs to the computation-intensive programming.Fast algorithms are needed in the image processing because it is the main cause of the system time bottleneck.In this paper,a number of classical filtering algorithms are studied,such as mean filtering,Gaussian filtering,median filtering and bilateral filtering.By comparing and optimizing the performance,the ICM software system can provide real-time or nearly real-time interaction for better user experience.