- Dona Paula, Goa, India.
- +91-0832- 2450327
- iiosc2020[at]nio[dot]org
| Abstract Submission No. | ABS-2022-14-0359 |
| Title of Abstract | Operational estimation of Radius of wind maximum (Rmax) during the Cyclone- Digital Image Processing Approach |
| Authors | SRINIVASA RAO N*, Gyaneshwar G |
| Organisation | INCOIS |
| Address | INCOIS K.V.Rangareddy, Telangana, India Pincode: 500090 Mobile: 9392011134 E-mail: srinivasn@incois.gov.in |
| Country | India |
| Presentation | Oral |
| Abstract | Cyclones are one of the natural disasters and catastrophic weather events on Earth. Over the years, the intensity and frequency of occurrence of tropical cyclone activity have significantly increased for various geophysical dynamics, especially in regions around India. Thanks to the meteorological satellite observations, which provide the necessary data to analyze and forecast the dynamics of cyclones. So, the task of tropical cyclone surveillance using remotely sensed data from satellite sensors is vital for disaster risk management. For instance, the estimation of cyclone centre or eye and radius of maximum wind (R_max) are the essential parameters for the forecast of cyclone dynamics and intensity, even for post-cyclone assessment analysis. However, most of the analyst's work to assess and forecast cyclone kinetics is still done manually, where a supervisor must perform the tasks in a static environment. In this regard, we develop a novel automatic segmentation algorithm to localize the cyclone region and center in thermal images using prominent analysis techniques. To the best of our knowledge, the proposed algorithm is the first of a kind to detect a cyclone region that mimics the human decision automatically. For these purposes, we build an algorithm that uses several artificial intelligence (AI) techniques from digital image processing and pattern recognition. The AI techniques in image processing and pattern recognition, such as adaptive thresholding, morphological operations, structure symmetry analysis, pattern matching, and isothermal contour analysis, helps to perform object recognition, segmentation, tracking of cyclone regions, and estimation of R_max. To assess and evaluate the algorithm's performance, we conducted various experiments using the cloud-top temperature (CTT) images with varying complexity of cyclones structure. The CTT images (from 2007 to 2021) are from an advanced very high-resolution radiometer (AVHRR) sensor from different satellites such as Metop-1, Metop-2, NOAA 18, and NOAA 19. The obtained cyclone region identification and centre detection results demonstrate the potential and feasibility of the proposed algorithm in a complete operational environment for the first time. The proposed algorithm provides a thoughtful perspective on automatic cyclone recognition, R_max estimation, and track determination that opens many innovative attitudes towards analyzing cyclones using satellite imagery. |