每年出版 12 期
ISSN 打印: 0040-2508
ISSN 在线: 1943-6009
Indexed in
Wavelet Coefficient Thresholding for Radar Image Compression
摘要
In this paper several well-known thresholding techniques: hard, soft, quantile and universal are represented. A new adaptive thresholding technique is proposed. It is known that radar images are affected by the phenomenon of multiplicative noise. This phenomenon influences the performance of 2D wavelet transform based compression algorithms and it can be improved if suitable algorithm is used as a preprocessing stage. It is shown that the homomorphic transform can be used as a suitable preprocessing technique. Comparison between well-known thresholding techniques and the proposed adaptive thresholding technique is performed for the real radar image. It is shown that the adaptive thresholding use improves the performance of the wavelet transform based compression.