Research Article
Spatiotemporal Super-Resolution Reconstruction Based on Robust Optical Flow and Zernike Moment for Video Sequences
Table 3
Average PSNR, MSSIM, and RMSE values in the four SR reconstruction methods.
| Video sequence | Evaluation index | SR reconstruction methods | POCS | NL-SR | ZM-SR | STSR (proposed) |
| Satellite-1 | PSNR (/dB) | 31.5435 | 32.2815 | 32.5435 | 34.2740 | MSSIM | 0.7621 | 0.7747 | 0.8032 | 0.8530 | RMSE | 0.0054 | 0.0056 | 0.0049 | 0.0032 |
| Satellite-2 | PSNR (/dB) | 20.5592 | 27.5057 | 28.6850 | 32.1723 | MSSIM | 0.8915 | 0.8979 | 0.9072 | 0.9483 | RMSE | 0.0146 | 0.0079 | 0.0071 | 0.0051 |
| Forman | PSNR (/dB) | 34.5542 | 35.1881 | 35.3928 | 36.6380 | MSSIM | 0.6533 | 0.8051 | 0.8171 | 0.8491 | RMSE | 0.0054 | 0.0049 | 0.0046 | 0.0038 |
| Suzie | PSNR (/dB) | 41.1444 | 44.3360 | 44.7565 | 45.4872 | MSSIM | 0.7031 | 0.7974 | 0.8006 | 0.8236 | RMSE | 0.0077 | 0.0055 | 0.0054 | 0.0046 |
|
|