Research And Implementation Of Image Copy-move Forgery Detection

Digital image is one of the most important media for disseminating information in the current society,thus various kinds of image processing software are bred,this kind of software is simple and powerful,ordinary users can easily modify images by using this kind of software,which brings convenience to users.however,image tampering for the purpose of distorting facts will bring great losses or adverse effects to people.Therefore,image forensics technology aiming that whether image has been tampered with has gradually become a research hotspot in the field of digital image.Copy-move tampering of the same image copies and pastes a target area into different positions of the same image,the tampered image will have many similar regions.Copy-move tampering detection technology of the same image belongs to image authenticity forensics technology in image passive forensics technology,it mainly identifies whether the image has multiple similar regions.This thesis mainly studies the same image copy-move tampering and detection technology,the main work includes:1.Researching the same image copy-move tampering and detection technology,including the study of common digital image tampering methods,implementation principles and the implementation process of various classical algorithms;researching two important processes of feature extraction and feature matching in the same image copy-move tampering detection process;and the contents of the research are simulated.2.Aiming at the problem that block matching detection algorithm extracts features of the gray image directly from the image to be detected,and the detection effect is poor,beside,natural similar regions and smooth regions of the image will lead to false detection,a method based on HSV color space and histogram equalization algorithm for accurate detection and location of copy-move tampering is proposed.Firstly,the image to be measured is transformed from RGB color space to HSV color space,and the V channel of HSV color space is enhanced by histogram equalization algorithm;secondly,the image is divided into overlapping image blocks,and the feature of H,S and V components are extracted to construct the feature vectors.then,the nearest neighbor search algorithm is used to search the constructed feature vectors quickly and randomly,and the dense migration field is obtained;finally,the dense linear fitting methodeliminates mismatches and locates copy-move areas by morphological operation marks.It can be proved by experiments that this method has good detection effect and can reduce the false detection pixels in natural similar areas and smooth areas,besides which can locate the copied and moved area of the image more accurately.3.An image copy-move tampering detection system is designed and implemented using GUI tool in software MATLAB.The system can detect whether the image has been tampered with by copy-move forgery,at the same time,the system can locate and identify the copy-move areas of the image more accurately.