CAJ | 학술논문

针对常用的拼接缝消除方法在图像间灰度差较大的情况下适用性差，甚至会导致新灰度突变的问题，提出一种新的拼接缝消除方法。首先以拼接缝两侧相邻像素的改正比值为基准比值，再将其分别在改正范围内逐渐过渡得到其他像素的改正比值，然后沿拼接缝方向对改正比值进行卷积平滑以避免条纹状的改正痕迹现象，最后与原图像加权融合得到无缝的结果图像。实验结果表明，算法既实现灰度的平滑过渡又较好地保持了原图像的灰度特征，与原图像的相似度和吻合度更高，结果图像具有良好的视觉效果和较高的图像质量，尤其对灰度差增大的情况仍能得到令人满意的结果。此外，根据拼接缝两侧平均灰度差值提出一种改正范围与灰度差成正比的自动确定改正宽度的方法，算法可行、有效，具有较强的适用性。
침대상용적병접봉소제방법재도상간회도차교대적정황하괄용성차，심지회도치신회도돌변적문제，제출일충신적병접봉소제방법。수선이병접봉량측상린상소적개정비치위기준비치，재장기분별재개정범위내축점과도득도기타상소적개정비치，연후연병접봉방향대개정비치진행권적평활이피면조문상적개정흔적현상，최후여원도상가권융합득도무봉적결과도상。실험결과표명，산법기실현회도적평활과도우교호지보지료원도상적회도특정，여원도상적상사도화문합도경고，결과도상구유량호적시각효과화교고적도상질량，우기대회도차증대적정황잉능득도령인만의적결과。차외，근거병접봉량측평균회도차치제출일충개정범위여회도차성정비적자동학정개정관도적방법，산법가행、유효，구유교강적괄용성。
The conventional seamline removal methods are not effective when the gray difference is great,which may cause new gray mutations .To solve the problem,we put forward a novel seamline removal method.Firstly,the gray correction ratio of the adjoined pixels was used as the standard correction ratio .And then the correction ratio of other pixels was obtained by making it change gradually .Meanwhile,convolution smoothing was taken to avoid the stripe correction trace .The image without any seamline was finally obtained after weighted fusion with the original image .Experiment results indicate that seamline is eliminated smoothly and the gray information is kept as well .The similarity and coincidence are higher and the result image has a good visual effect and high quality .The effect of seamline removal is better especially when the gray difference is great .Apart from that,an automatic correction width confirming method was proposed,which is proportional to the average gray difference between the two sides of the seam -line.The proposed method is applicable,effective and easy to realize .