在有限脊波变换的基础上,根据对图像有限Radon变换系数的分析,提出了一种自适应的正交有限脊波变换。此变换根据特别设计的代价函数,可自适应地为每列有限Radon变换系数选择最优的正交变换形式,从而可高效地分析图像中的直线奇异。讨论了此变换在图像编码领域中的应用,通过对零树编码算法SPIHT进行改进,实现了图像自适应正交有限脊波变换系数的渐进编码。实验结果表明,对富含直线边缘的图像来说,与小波图像编码算法相比,该算法获得了较高的边缘重建质量。 Based on the finite ridgelet transform, an adaptive orthogonal finite ridgelet transform is proposed based on the analysis of the limited Radon transform coefficients. Based on a specially designed cost function, this transform adaptively selects the optimal orthogonal transform for each of the finite Radon transform coefficients in each column, allowing for efficient analysis of linear singularities in the image. This paper discusses the application of this transform in the field of image coding. By improving SPIHT algorithm, progressive coding of image adaptive orthogonal ridgelet transform coefficients is achieved. Experimental results show that the proposed algorithm has higher edge reconstruction quality than the wavelet image coding algorithm for images with linear edges.