Aiming to improve the quality of JPEG decompressed infrared images, this paper presents a constrained optimization method for blocking artifact compensation. First, we define the drop value between image blocks and measure the DC component loss by the drop value. Second, a constrained optimization object function is constructed based on the DC component compensation model. Both global and local constraints are used to avoid overcompensation. Finally, the optimal compensation value is obtained using iterative processing. Experimental results show that the proposed algorithm can remove the blocking artifact in JPEG infrared image effectively and outperform the existing methods.

为了提高JPEG解压缩红外图像的质量，提出了一种用于块伪影补偿的约束优化方法。首先，我们定义图像块之间的下降值，并通过下降值测量直流分量损失。其次，基于直流分量补偿模型构建约束优化目标函数。全局和局部约束都用于避免过度补偿。最后，通过迭代处理得到最优补偿值。实验结果表明，该算法能有效去除JPEG红外图像中的块状伪影，优于现有方法。