【摘要】：隨著圖像處理技術(shù)的迅速發(fā)展,對數(shù)字圖像(自然圖像)內(nèi)容的篡改變得越來越容易,各類虛假圖像頻繁出現(xiàn),使得數(shù)字圖像可靠性,穩(wěn)健性遭到嚴(yán)重?fù)p害。因?yàn)楹芏鄶?shù)字圖像攝取設(shè)備并不具備嵌入數(shù)字水印功能,所以基于數(shù)字水印的主動取證技術(shù)受到很大限制,數(shù)字圖像盲取證技術(shù)因此應(yīng)運(yùn)而生。數(shù)字圖像篡改盲取證是在對原始圖像沒有任何先驗(yàn)知識,不對原始圖像進(jìn)行任何事先處理的情況下,對數(shù)字圖像從被數(shù)字相機(jī)拍攝以后沒有經(jīng)過篡改的取證。數(shù)字圖像盲取證技術(shù)主要應(yīng)用于檢測法庭上作證的照片、媒體上出現(xiàn)的照片是否真實(shí)可靠,還原出模糊退化圖像的內(nèi)容作為法庭證據(jù)等。這是一門剛剛興起的技術(shù),隨著數(shù)字圖像越來越廣泛的應(yīng)用,數(shù)字圖像盲取證技術(shù)也會有越來越廣闊的應(yīng)用前景。本文主要通過對圖像篡改過程中拼接痕跡的檢測技術(shù),實(shí)現(xiàn)了兩種檢測圖像是否被篡改的方法。第一種方法利用圖像塊之間的特征值來進(jìn)行相互比較,并判別出同幅圖像中相似的圖塊,此部分即可判斷為經(jīng)過復(fù)制粘貼篡改的部分。第二種檢測方法是基于圖像邊緣模糊篡改檢測,利用一個濾波器實(shí)現(xiàn)對圖像的邊緣模糊的檢測。經(jīng)過實(shí)驗(yàn)證明,以上兩種方法均能成功檢測出同一幅圖像中的復(fù)制粘貼篡改部分。
[Abstract]:With the rapid development of image processing technology, the tampering of digital image (natural image) has become more and more easy, and various kinds of false images appear frequently, which seriously damages the reliability and robustness of digital image. Because many digital image acquisition devices do not have the function of embedding digital watermarking, the active forensics technology based on digital watermarking is greatly restricted, so digital image blind forensics technology emerges as the times require. Blind forensics of digital image tampering is to obtain evidence without any prior knowledge of the original image and any prior processing of the original image, and the digital image has not been tampered with since it was taken by the digital camera. Blind digital image forensics technology is mainly used to detect the evidence in court, whether the photos appear in the media are true and reliable, and to restore the content of the blurred degraded image as court evidence. This is a newly emerging technology. With the more and more extensive application of digital image, blind forensics of digital image will have more and more broad application prospects. In this paper, two methods of detecting whether the image is tampered with or not are realized by detecting the stitching marks in the process of image tampering. The first method makes use of the eigenvalues of the image blocks to compare with each other and to distinguish the similar blocks in the same image. This part can be judged as the part that has been tampered with by copy and paste. The second method is based on image edge fuzzy tamper detection, using a filter to detect image edge blur. The experiments show that the two methods can detect the copy and paste tampering part of the same image successfully.
【學(xué)位授予單位】：天津大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2016
【分類號】：TP391.41

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 何碧連;;數(shù)字圖像資源的檢索與服務(wù)[J];圖書館研究與工作;2008年03期

2 王玨;;數(shù)字圖像質(zhì)量解析[J];美術(shù)大觀;2010年01期

3 王剛;;數(shù)字圖像的快速與科學(xué)識別研究[J];藝術(shù)與設(shè)計(理論);2012年12期

4 黃];楊h接，

本文編號：2407087