稀疏编码就是对人类等哺乳动物视觉系统主视皮层强大图像编码能力的成功模拟，具有自适应性，且得到的图像基具有空间的局部性、方向性和频域的带通性。在稀疏编码基础上发展而来的非负稀疏编码，克服了特征间的相互抵消现象，编码性能更为优越。而利用经验模态分解技术加入图像结构信息的非负稀疏编码方法，在兼顾非负稀疏编码特性的基础上能更好地体现图像的结构性特征。本文提出了基于图像基的图像压缩方法，把这种改进的非负稀疏编码算法用于图像压缩，在保证较好图像解码质量的情况下，获得了理想的压缩比。

In the long process of human evolution and development,every day,the human visual system deals with a lot of visual information. For its own survival and development,the human beings have to evolve a powerful imageprocessing capability,such as image coding of the visual system,etc. Sparse coding successfully simulats the strong ability of image coding of the primary visual cortex in mammals’ visual systems such as human,which features selfadaptability,and the learned image base are localized,oriented in space domain and bandpass in the frequency domain. Nonnegative sparse coding developes and evolves on the basis of sparse coding,to overcome the interoffset phenomena of characteristics,and the coding performance is more superior. And the nonnegative sparse coding,which joins the  image structure information using the experience modality decomposition technology,reflects the structural characteristics of the image better in addition to ensuring the characteristics of nonnegative sparse coding. An image coding approach based on the image base is presented in this paper,and the improved nonnegative sparse coding algorithm is used in image compression,and an ideal compression ratio in ensuring a better quality of the decoded images is achieved.