Improved EZW Compression Coding Algorithm for Coal Gangue Image

In the coal gangue automatic selection system, many coal gangue images have been created. It is urgent problem to store and transfer these coal gangue images in real-time processing. An improved embedded zerotree wavelet (EZW) coding algorithm is introduced to compress the coal gangue images in this paper. According to wavelet transform theory, the basic features of wavelet coefficients are analyzed. The zerotrees are divided into the edge tree, texture tree and smoothness tree, and the different coding schemes are applied to these different trees. The low frequency components in the coal gangue images are compressed in lossless, whereas the wavelet coefficients of the high frequency in different directions are processed with the adaptive threshold values. The integer wavelet transform is adopted to keep the higher fidelity in inversion operation. Investigation results show that the proposed algorithm can raise the compression efficiency and gain the higher human vision quality in the reconstructed images.