MACHINE LEARNING BASED CROP CYCLE MAPPING USING MULTI TEMPORAL SPACE DATASETS

Gupta, P. K.; Verma, P. A.; Chauhan, P.

The Indo-Gangetic region is an important source of food grain produced in India. With advent of advanced crop management practices, there has been a recent increase in production. This could be due to improved yield and/or multiple crops per year. Understanding cropping patterns is, therefore, a fundamental requirement to strategize food security in the country. The long running space based earth observation programmes, provides a powerful tool to study the change in vegetation, through use of indices, synoptically and temporally. This study proposes a machine learning based technique to identify distributed (pixel-wise) cropping cycles, based on the temporal trend in normalized difference vegetation index (NDVI). The technique is applied to several time series NDVI data sets (OCM, MODIS and GIMMS) and results are discussed. There is tremendous scope to further extend this study to predict drought conditions, change in cropping pattern, and estimate production statistics, by using advanced artificial intelligence methods (such as deep learning).

Zitieren

Zitierform:

Gupta, P. K. / Verma, P. A. / Chauhan, P.: MACHINE LEARNING BASED CROP CYCLE MAPPING USING MULTI TEMPORAL SPACE DATASETS. 2019. Copernicus Publications.

Rechte

Rechteinhaber: P. K. Gupta et al.

Nutzung und Vervielfältigung:

Export