中文摘要：

以QuickBird卫星记录的DN值图像为实验数据,根据1～4波段积分辐亮度之和与全色波段积分辐亮度之间的比例关系及能量守恒律构建分解方程,将1～4波段图像每个像元分解成16（4×4）个像元,获得了分辨率为0.61m的1～4波段辐亮度图像。与原图像比较,空间分辨率提高到4倍,地物细部特征清晰,视觉效果明显提高,图像成图适宜比例尺从1∶10000提高到1∶2500。

英文摘要：

In the condition that the DN（Digital Number） value images of ground-atmosphere system recorded by the QuickBird artificial satellite sensor were taken as the experimental datum,the equation of pixel decomposition was structured,abiding by energy conservation law and the proportional relationship between the sum of 1-4 band integral radiance and the panchromatic band （Pan） integral radiance. By using the equation gained above,every pixel of 1-4 band images was decomposed into sixteen pixels. So the high resolution 1-4 band radiance images with the resolution of 0.61 m were finally obtained in this experiment. Compared with the original 1-4 band images,the spatial resolution of the high resolution 1-4 band radiance images was raised by four times. The detailed characteristics of the surface features in the high resolution 1-4 band radiance images were more limpid than the original 1-4 band images. The optesthesia effect was also improved obviously in the high resolution 1-4 band radiance images. Therefore,the map with appropriate scale improved from 1∶10 000 to 1∶2 500 can be made based on the high resolution 1-4 band radiance images.