CAJ | 학술논문

针对全景泊车辅助系统中图像融合效果较差的问题，提出一种渐入渐出的图像融合算法。首先，根据配准完成的全景图像选取出重合区域，并确定其边界的坐标参数值；在此基础上，计算该区域中像素点坐标到其边界的最短距离以得出融合算法的权值，并采用上述权值进行图像融合，从而得出汽车四周的全景图像；最后，利用实例验证了该方法的有效性与可行性。
침대전경박차보조계통중도상융합효과교차적문제，제출일충점입점출적도상융합산법。수선，근거배준완성적전경도상선취출중합구역，병학정기변계적좌표삼수치；재차기출상，계산해구역중상소점좌표도기변계적최단거리이득출융합산법적권치，병채용상술권치진행도상융합，종이득출기차사주적전경도상；최후，이용실례험증료해방법적유효성여가행성。
The issue of fade-in-fade-out algorithm for panoramic image fusion of parking assist system was proposed to improve the visual effect of image fusion.The overlap region was determined by accord-ing to panoramic image of registration.And the coordinates of the boundary parameter were determine. Based on this,the shortest distance of the pixel coordinates to its borders was derived to determine the weighting of fusion algorithm.And these weights of image fusion were used.Then the panoramic image a-round the car was given.Finally,the examples were presented to verify the feasibility and effectiveness of the proposed results.