X-ray inspections need good quality images so that features in the material can be observed. The Wavelet Image Spectra Enhancement (WISE) technique was developed as a set of powerful filters designed to improve the quality of x-ray images. The overall quality of these images is critical to a successful x-ray inspection.
In this paper we describe how the transform thresholding and parameter estimation techniques in WISE are used as building blocks for radiographic feature augmentation. Transform thresholding techniques are applied to select the frequency coefficients of the discrete wavelet transform (DWT), the discrete cosine transform (DCT), and the discrete Fourier transform (DFT). Furthermore, a novel algorithm using the continuous wavelet transform (CWT) is also presented as a parameter estimation approach to filter x-ray imaging data. The data filtering performance of these algorithms is examined using both simulated and experimental x-ray signals. The results show that the parameter estimation has superior filtering performance when compared to the conventional transform thresholding techniques.
X-ray inspection greatly benefits from sharpening and noise suppression algorithms. The data augmentation performance of the DWT, the DFT, and the DCT are examined when applied to simulated and experimental x-ray data. An adaptive threshold is applied to remove the smaller coefficients of the frequency domains to achieve data filtering. Then, the signals are reconstructed from the remaining coefficients. The mean square errors of these techniques are analyzed for filtering and compression ratios of up to 90%. For higher ratios, the CWT is also investigated in this study. The CWT performs the correlation of a kernel wavelet with the radiographic signal. A modified version of the Morlet wavelet is used as the kernel wavelet to estimate the signal parameters. Upon the determination of these parameters, high data improvement and denoising can be obtained. The CWT is applied to both simulated and experimental data, and the results are presented in this paper.
Please download here a copy of the white paper “WISE: WAVELET IMAGE SPECTRA ENHANCEMENT OF X-RAY IMAGES”.