An Infrared Image Enhancement Technique Based on Neighborhood Wavelet Thresholding Coefficient for Multi-level Discrete Wavelet Transform

The quality of infrared images is critical to a wide range of emerging applications and research. These images however suffer from low contrast and poor image quality. This research proposes an infrared image enhancement technique based on neighborhood wavelet thresholding coefficients for multi-level discrete wavelet transform (DWT). This technique will be implemented by first performing image pre-processing using Dabeuchies D4 filter. Following image preprocessing is multi-level wavelet decomposition based on the proposed neighborhood thresholding technique. Based on this thresholding technique, noise is eliminated and the sub-images undergo multi-level wavelet reconstruction to give an enhanced image as the final output. The results obtained were subjected to quantitative assessment by computing the peak signal-to-noise-ratio (PSNR) and discrete entropy (DE) values. From the assessment, the developed technique significantly eliminates noise with a better dynamic range.