破损图像丢失区域自适应优化修复方法仿真
|
摘要
针对传统修复方法存在图像模糊、修复速度低的问题,提出一种基于纹理合成的破损图像丢失区域自适应修复方法。对破损图像丢失区域进行纹理分割提取破损图像边缘能量特征,完成自适应纹理密度量化估计,计算丢失区域纹理向量量化区域的超像素级视觉特征。利用基于Criminisi算法根据纹理特性计算优先修复区域,不断更新图像的可靠度对丢失区域进行纹理修复,完成破损图像丢失区域的自适应修复。通过实验结果验证,所提修复方法与传统修复方法相比,修复后的图像更清晰、完整,修复用时更短。
In traditional restoration methods, the image is fuzzy and restoration speed is low. Therefore, an adaptive restoration method for lost region in damaged image based on texture synthesis is proposed. At first, the energy feature of damaged image edge was extracted by texture segmentation of lost region in damaged image. Then, the adaptive texture density quantization estimation was completed. The texture vector of lost area was calculated and super-pixel visual feature was quantized. Based on Criminisi algorithm, priority restoration region was computed through the texture characteristics. By constantly updating the reliability of image, the lost regions were repaired. Thus, the adaptive restoration of lost regions in damaged image was completed. Simulation results show that, compared with traditional method, the restored image is clearer and more complete through the proposed restoration method. Meanwhile, the repair time is shorter.
In traditional restoration methods, the image is fuzzy and restoration speed is low. Therefore, an adaptive restoration method for lost region in damaged image based on texture synthesis is proposed. At first, the energy feature of damaged image edge was extracted by texture segmentation of lost region in damaged image. Then, the adaptive texture density quantization estimation was completed. The texture vector of lost area was calculated and super-pixel visual feature was quantized. Based on Criminisi algorithm, priority restoration region was computed through the texture characteristics. By constantly updating the reliability of image, the lost regions were repaired. Thus, the adaptive restoration of lost regions in damaged image was completed. Simulation results show that, compared with traditional method, the restored image is clearer and more complete through the proposed restoration method. Meanwhile, the repair time is shorter.
引文
[1] 李旭峰,王静,刘红敏,等.特征优先块匹配图像修复算法[J].计算机辅助设计与图形学学报,2016,28(7):1131-1137.
[2] 毛宇航,李光耀,肖莽.基于平面结构信息的图像修复[J].计算机工程,2016,42(3):236-241.
[3] 王斌,等.基于小波域稀疏最优的图像修复方法[J].电子学报,2016,44(3):600-606.
[4] 廖斌,苏涛.基于窄带优化的自适应多匹配块随机查找图像修复[J].量子电子学报,2017,34(6):656-661.
[5] 段维夏,聂洪玉,王猛.基于破损区域分类的自适应扩散模型[J].计算机应用研究,2016,33(9):2823-2826.
[6] 吴银芳,朱森诚.基于匹配调节法则和梯度约束模型的图像修复算法[J].包装工程,2018,39(13):239-244.
[7] 刘华明,毕学慧,叶中付,等.样本块搜索和优先权填充的弧形推进图像修复[J].中国图象图形学报,2016,21(8):993-1003.
[8] 沈跃,徐慧,刘慧,等.基于K-means和近邻回归算法的Kinect植株深度图像修复[J].农业工程学报,2016,32(19):188-194.
[9] 许有俊,等.土压平衡顶管变形破损修复仿真研究[J].计算机仿真,2017,34(10):237-240.
[10] 何仕文,刘琳,张永强,等.改进TV-H～(-1)模型的图像修复方法[J].哈尔滨工业大学学报,2016,48(2):167-172.
[2] 毛宇航,李光耀,肖莽.基于平面结构信息的图像修复[J].计算机工程,2016,42(3):236-241.
[3] 王斌,等.基于小波域稀疏最优的图像修复方法[J].电子学报,2016,44(3):600-606.
[4] 廖斌,苏涛.基于窄带优化的自适应多匹配块随机查找图像修复[J].量子电子学报,2017,34(6):656-661.
[5] 段维夏,聂洪玉,王猛.基于破损区域分类的自适应扩散模型[J].计算机应用研究,2016,33(9):2823-2826.
[6] 吴银芳,朱森诚.基于匹配调节法则和梯度约束模型的图像修复算法[J].包装工程,2018,39(13):239-244.
[7] 刘华明,毕学慧,叶中付,等.样本块搜索和优先权填充的弧形推进图像修复[J].中国图象图形学报,2016,21(8):993-1003.
[8] 沈跃,徐慧,刘慧,等.基于K-means和近邻回归算法的Kinect植株深度图像修复[J].农业工程学报,2016,32(19):188-194.
[9] 许有俊,等.土压平衡顶管变形破损修复仿真研究[J].计算机仿真,2017,34(10):237-240.
[10] 何仕文,刘琳,张永强,等.改进TV-H～(-1)模型的图像修复方法[J].哈尔滨工业大学学报,2016,48(2):167-172.