A new satellite (code:03) was launched on October 26, 2015. It belongs to Mapping Satellite-1 (TH-1), which is a group of High-Resolution (HR) optical stereo mapping satellites in China. Three-linear-array/multispectral (TLA/MUX) images are the important data products of TH-1.They are used for stereo mapping, fusion and classification, and remote sensing. Image quality assessment is a critical issue particularly in the regular operation of the HR satellite ground application system, with the assessment results having great significance in the application prospects of the current satellite and subsequent satellite development.
In this study, a multi-objective index method is employed to evaluate comprehensive the image quality of the TH-1 03 TLA/MUX sensors. It adopts 10 objective indices, including clarity (CLA), contrast(CON), detail energy(DET), and edge energy (EDG).Given that the three in-orbit MS-1 satellites have the same sensor and orbit parameters, the images from TH-1 01 and TH-1 02 are used as comparison assessment data. A group of images with the same screening season, terrain type, and natural environment are separately selected from the three in-orbit MS-1 satellites to avoid assessment errors induced by differences in seasons, terrain types, and natural environments.
Assessment results show that the objective indices of the TH-1 03 TLA/MUX images, including CLA, CON, DET, EDG, power spectral density, and information content are significantly higher than those of the images of TH-1 01 and TH-1 02. The higher values of these indices demonstrate that the TLA/MUX images of TH-1 03 are better than TH-1 01 and TH-1 02 satellites in terms of describing the detail texture and edge texture features of ground objects. The TLA/MUX images of TH-1 03 have higher information content as well. The signal-to-noise ratio reveals that the capability of the TLA/MUX sensors of MS-1-03 to suppress noise is superior to those of the sensors of TH-1 01 and TH-1 02. Furthermore, RadEdge and GainAdj indicate that the radiometric uniform level of TH-1 03 is in between those of the TH-1 01 and TH-1 02 satellites.
The results of the image quality assessment for TH-1 03 TLA/MUX sensors are analyzed and reported for the first time, and the preliminary conclusion is discussed in this paper. According to the assessment results, the image quality of TLA/MUX sensors of TH-1 03 is improved compared with those of the TLA/MUX sensors of the TH-1 01 and TH-1 02 satellites. Owing to the development in the image quality, the use of TLA/MUX data of the TH-1 03 satellite can enhance target recognition, classification, and stereo mapping accuracy. In summary, this paper has presented the first study on the analysis of the image quality for TH-1 03 TLA/MUX sensors and has made an exemplary contribution.