Abstract—In the present world, predicting rainfall is considered to be an essential and also a challenging task. Normally, the climate and rainfall are presumed to have non-linear as well as intricate phenomena. For predicting accurate rainfall, we necessitate advanced computer modeling and simulation. When there is an enhanced understanding of the spatial and temporal distribution of precipitation then it becomes enrichment to applications such as hydrologic, climatic and ecological. Conversely, there may be some kind of challenges occur in the community due to some application which results in the absence of consistent precipitation observation in remote and also emerging region. This survey paper provides a multifarious collection of methodologies which are epitomized by various researchers for predicting the rainfall. It also gives information about some technique to forecast rainfall, which is appropriate to all methods like numerical, traditional and statistical. Keywords—Satellite Image, Segmentation, Feature Extraction, Classification, Clustering, Precipitation Estimation. I. INTRODUCTION AINFALL carries the supreme vital role in the matter of human life in entire manners of weather conditions. For human cultivation the influence of rainfall is very massive. Rainfall is considered to be one of the most natural climatic wonders whose prediction is arduous and challenging. The exact information about to bring rainfall plays a main role in the development and management of water assets and similarly important for prevention from reservoir maneuver and floods. In the metropolitan areas, rainfall has a durable impact on traffic, sewer systems and also some more human undertakings. On the other hand, the hydrology cycle of rainfall is one of the most composite and problematic elements to recognize and to model. This is due to the complexity of the atmospheric processes which create rainfall and the significant range of variation over a wide range of scales mutually in time and space. In recent epochs, predicting the accurate rainfall seems to be extreme challenges in operational hydrology. Rainfall predicting is closely associated with the agricultural region, whereas in their term rainfall means crops and crops means life. Agriculture plays an important role to enhance the economy of the nation. By using diverse methods; the huge Mrs.S.Sarumathi, Associate Professor, is with the Department of Information Technology, K. S. Rangasamy College of Technology, Tamil Nadu, India (phone: 9443321692; e-mail: rishi_saru20@rediffmail.com). Dr. N. Shanthi, Professor and Dean, is with the Department of Computer Science Engineering, Nandha Engineering College, Tamil Nadu, India (e- mail: shanthimoorthi@yahoo.com). Ms. S. Vidhya, PG Scholar, is with the Department of Information Technology, K. S. Rangasamy College of Technology, Tamil Nadu, India (phone: 9443960666; e-mail: vidhyapsubramani@gmail.com). scale of efforts has been undertaken by various researchers and scientist for predicting the rainfall accurately and effectively over the universe. However the accurate rainfall prediction estimated by several techniques was not fully satisfied still now because the rainfall has nonlinear nature [1]. For the past era, there are many satellite sensor technology has simplified for the growth of innovative methods to global precipitation observations. In recent times various satellite- based precipitation algorithms have been established which produce precipitation products involving of higher spatial and temporal resolution that is useful for hydrologic researches and water resources applications [2]. II. DIFFERENT SATELLITE PREDICTION TECHNIQUES A. Precipitation Estimation from Remotely Sensed Information Using Neural Networks (PERSIANN) An automated system for Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks, the PERSIANN [3] has been established. This is mainly constructed for estimating the rainfall from geosynchronous satellite longwave infrared imagery. The Tropical Rainfall Measurement Mission (TRMM) data are responsible for large-scale estimates of tropical rainfall over the long term. The estimation of surface rainfall can be reported by several algorithms which use geostationary longwave infrared channel images (e.g., [4], [5]). The realistic high spatial and temporal resolution (~4 ൈ 4 Km ଶ every 30 min) can be undertaken by the measurements with wide coverage of the land turned over by the geosynchronous satellites which is taken to be the strong point of this overture. On behalf of monitoring the spatial and temporal development of clouds these measurements can be utilized, the measurements of cloud-top brightness temperatures are offered by the longwave infrared (IR) channels that they do not supply enough information to meet the real volume of rain occurring at the soil surface. Depend upon the multispectral microwave measurements made by polar-orbiting satellites several algorithms have been made to calculate approximately rainfall. The multispectral microwave sensors have the ability to penetrate into the clouds and therefore within the hydrometeor column the measured brightness temperature depends on the emission, absorption process. The building of the hydrometeor column can be obtained by using radiative transfer models of the emission-absorption process and rapid rain rates can be estimated by physically based algorithms. Shortly from both geosynchronous and polar-orbiting satellites, the study has been undertaken for the development of methods which abuse the assets of a variety of sensors. S. Sarumathi, N. Shanthi, S. Vidhya Satellite Rainfall Prediction Techniques - A State of the Art Review R World Academy of Science, Engineering and Technology International Journal of Vol:9, No:2, 2015 144 International Scholarly and Scientific Research & Innovation 9(2) 2015 scholar.waset.org/1307-6892/10001177 International Science Index, Vol:9, No:2, 2015 waset.org/Publication/10001177