CAJ | 학술논문

焊接数字图像因其易被篡改的特性而影响其完整性，对焊接质量的评价过程造成很大困难，从而影响评价结果。文章建立原始焊接数字图像的敏感奇异值组，基于二维混沌序列加密水印图像，将其嵌入至原始焊接数字图像中，通过对比提取出的水印完成篡改的检测和定位，并对实验进行了验证。结果表明：算法峰值信噪比为84.29，大于等于36 dB，能够视觉保真；算法的篡改检出概率不小于99.99946％；经过对100个篡改样本进行试验，水印相似度均小于1，算法可成功完成对篡改样本检出和定位。
한접수자도상인기역피찬개적특성이영향기완정성，대한접질량적평개과정조성흔대곤난，종이영향평개결과。문장건립원시한접수자도상적민감기이치조，기우이유혼돈서렬가밀수인도상，장기감입지원시한접수자도상중，통과대비제취출적수인완성찬개적검측화정위，병대실험진행료험증。결과표명：산법봉치신조비위84.29，대우등우36 dB，능구시각보진；산법적찬개검출개솔불소우99.99946％；경과대100개찬개양본진행시험，수인상사도균소우1，산법가성공완성대찬개양본검출화정위。
The integrity of the welding digital image is threatened due to its characteristics of being easy to be tampered and brings difficulties to the quality evaluation process and affects the evaluation result. This paper establishes the sensitive SVD group of the original welding digital image. Based on two-dimensional chaotic sequence,the watermark image is encrypted and embedded into the original welding digital image. By comparing the distracted watermark with the original one,the tamperings are detected and located. This algorithm is finally verified by experiments. The results show that the visual fidelity is assured by the peak signal-to-noise ratio which is greater than 36 dB. The detection probability is larger than 99. 99946 % . By the experiments on 100 tampered samples,the watermark similarities are all less than 1. This algorithm successfully detected and located the tampered samples.