Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM68789
Volume:
10 |
11 |
Nov 2026
Pages:
1-5
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Computer Engineering
Adsul's Technical Campus
Computer Engineering
Adsul's Technical Campus
Computer Engineering
Adsul's Technical Campus
Computer Engineering
Adsul's Technical Campus
Computer Engineering
Adsul's Technical Campus
Computer Engineering
Adsul's Technical Campus
Geostationary weather satellites such as GOES-16 and INSAT-3D usually deliver a full-disk image only every 10 to 30 minutes, which is too coarse for tracking fast convective storms. This paper presents a software-only system that converts 30-minute infrared satellite sequences into 5-minute sequences by generating intermediate frames with deep optical flow. Infrared frames from the SEVIR dataset, NOAA GOES-16/18 archives and INSAT-3D data are cleaned, resized to 256 × 256 pixels and normalised. A Real-Time Intermediate Flow Estimation (RIFE) network, fine-tuned on satellite frame sequences, predicts bidirectional intermediate flows and a fusion mask for any timestep, and the two input frames are backward-warped and fused to produce five new frames between each real pair. Because SEVIR provides real frames every 5 minutes, every generated frame can be scored against ground truth using PSNR, SSIM and MAE at 10, 20 and 30-minute input gaps. The proposed model is compared with linear blending, Farneback optical flow and Super SloMo under the same data split. Results and frame sequences are stored in MongoDB and served through a MERN web dashboard that plays real and generated frames side by side with clear labels.
KEYWORDS
This paper presented a software system that increases the temporal resolution of geostationary infrared imagery from 30 minutes to 5 minutes using a RIFE network fine-tuned on satellite frames. The pipeline covers reading NetCDF and HDF5 products from SEVIR, NOAA and MOSDAC, preprocessing, arbitrary-timestep interpolation, evaluation against real held-out frames at 10, 20 and 30-minute gaps, and a MERN dashboard that separates real and generated frames clearly. Because no new satellite hardware is needed, the approach can be applied to archived data as well as to near real-time feeds. Future work includes multi-band input combining water vapour and window channels, training directly on INSAT-3DR and INSAT-3DS rapid-scan data, adding physical constraints such as advection consistency to the loss, and uncertainty maps that show where the generated frames are least reliable.
1. Z. Huang, T. Zhang, W. Heng, B. Shi and S. Zhou, “Real-Time Intermediate Flow Estimation for Video Frame Interpolation,” in Proc. European Conference on Computer Vision (ECCV), 2022, pp. 624–642. 2. H. Jiang, D. Sun, V. Jampani, M.-H. Yang, E. Learned-Miller and J. Kautz, “Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation,” in Proc. IEEE/CVF CVPR, 2018, pp. 9000–9008. 3. M. S. Veillette, S. Samsi and C. J. Mattioli, “SEVIR: A Storm Event Imagery Dataset for Deep Learning Applications in Radar and Satellite Meteorology,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 33, 2020, pp. 22009–22019. 4. G. Farnebäck, “Two-Frame Motion Estimation Based on Polynomial Expansion,” in Proc. Scandinavian Conference on Image Analysis, LNCS vol. 2749, 2003, pp. 363–370. 5. B. K. P. Horn and B. G. Schunck, “Determining Optical Flow,” Artificial Intelligence, vol. 17, pp. 185–203, 1981. 6. Z. Wang, A. C. Bovik, H. R. Sheikh and E. P. Simoncelli, “Image Quality Assessment: From Error Visibility to Structural Similarity,” IEEE Transactions on Image Processing, vol. 13, no. 4, pp. 600–612, 2004. 7. T. J. Schmit, P. Griffith, M. M. Gunshor, J. M. Daniels, S. J. Goodman and W. J. Lebair, “A Closer Look at the ABI on the GOES-R Series,” Bulletin of the American Meteorological Society, vol. 98, no. 4, pp. 681–698, 2017. 8. A. Dosovitskiy et al., “FlowNet: Learning Optical Flow with Convolutional Networks,” in Proc. IEEE ICCV, 2015, pp. 2758–2766. 9. D. Sun, X. Yang, M.-Y. Liu and J. Kautz, “PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume,” in Proc. IEEE/CVF CVPR, 2018, pp. 8934–8943. 10. Z. Teed and J. Deng, “RAFT: Recurrent All-Pairs Field Transforms for Optical Flow,” in Proc. ECCV, 2020, pp. 402–419.
Publication Date
11/6/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
5
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...