Research Of Digital Image Watermarking Algorithm Based On Wavelet Transform And Spatial Domain

With the rapid development of Internet Technology, especially the great progress made in Web search engine technology, more and more digital products are illegally distributed on Internet in a much higher efficiency than ever before. Therefore, Digital Watermarking Technology, which was used in protecting copyrights, faced with a strong challenge. In recent years it has been highly concerned by IT industry and got in-depth research. It has a promising prospect.This paper studied the Static Gray image Watermarking technology used in copyright protection. First introduced basic theories about SGIW and its research situation, discussed the technologies used in digital image watermarking system, described the application of chaos theory, wavelet transform theory, HVS character in digital image watermarking technology. Based on analyses and conclusion of existing algorithms, we found many problems are not solved: most of the algorithms are not compatible with watermark because of watermark's invisibility, robustness and embedding capacity. Combined spatial domain algorithm with wavelet domain algorithms, this paper presents an improved combination domain algorithm. Based on the user's choice and the importance of data, depart an important part of from the watermark image, and embed this part into wavelet domain low frequency sub-band. According to the human visual system characteristics, the original image block can be divided into four parts: smooth zone, marginal zone, general zone and texture zone. Author achieved blind watermark detection by select low frequency coefficients of wavelet texture blocks as an embedded position, using of Parity Quantization Watermarking Algorithm. Finally, embedded the watermark image into spatial domain of the embedded image in wavelet domain, and embedding position is selected randomly by general cat map and logistic sequence. Watermark is embedded in the pixels 1 to 5 bits, but not the highest of non-zero bits.With the use of combination domain algorithm, even if the image was attacked and the watermark embedded in spatial domain was completely destroyed, the important watermark can still be obtained because wavelet domain algorithm's robustness. Thus the complete loss of watermark information can be avoided. Finally, we made a series of experiments such as compression attacking, noise attacking, filtering attacking and cropping attacking. Results turned out that this algorithm can improve the watermark capacity, with a good robustness and invisibility.Finally, the paper summarized the completed work, give prospects for research work next.