US 12,033,368 B2
Method and system for providing generalized approach for crop mapping across regions with varying characteristics
Jayantrao Mohite, Thane (IN); Suryakant Ashok Sawant, Thane (IN); Ankur Pandit, Thane (IN); and Srinivasu Pappula, Hyderabad (IN)
Assigned to Tata Consultancy Services Limited, Mumbai (IN)
Filed by Tata Consultancy Services Limited, Mumbai (IN)
Filed on Jan. 26, 2022, as Appl. No. 17/585,108.
Claims priority of application No. 202121003692 (IN), filed on Jan. 27, 2021.
Prior Publication US 2022/0237888 A1, Jul. 28, 2022
Int. Cl. G06V 10/48 (2022.01); G06F 18/214 (2023.01); G06N 20/00 (2019.01); G06V 10/74 (2022.01); G06V 10/774 (2022.01); G06V 20/10 (2022.01); G06V 20/13 (2022.01)
CPC G06V 10/48 (2022.01) [G06N 20/00 (2019.01); G06V 10/761 (2022.01); G06V 10/774 (2022.01)] 9 Claims
OG exemplary drawing
 
1. A processor implemented method for providing generalized approach for crop mapping across regions with varying characteristics, the method comprising:
generating, by one or more hardware processors, a labelled pixel dataset representing cropping pattern of a Region of Interest (ROI) for building a ML crop mapping model for the ROI, wherein the generated labelled pixel dataset captures regional dependency and localized phenological indicators for the ROI, and wherein the generating the labelled pixel dataset comprises:
receiving a plurality of remote sensing images of the ROI covering a plurality of climatic seasons spread across a predefined timespan;
randomly selecting a set of pixels from the plurality of remote sensing images and obtaining a plurality of pixel features for each of the plurality of pixels, wherein the plurality of pixel features comprise remote sensing and time series-based features predefined for the ROI;
analysing the plurality of pixel features obtained for each of the plurality of pixels with a plurality of crop features associated with each crop among a set of crops identified for the ROI, wherein the set of crops, and corresponding plurality of features for each of the set of crops are pre-identified for the ROI using Machine Learning (ML) based approach in accordance with agro-related historical information and localized phenological features obtained for the ROI from a plurality of sources; and
identifying and labelling a plurality of optimal pixels from the set of pixels with a crop among the set of crops to generate the labelled pixel dataset, wherein each of the plurality of optimal pixels are identified and labelled with an associated crop label based on degree of similarity computed between a) each of the plurality of optimal pixels and corresponding plurality of pixel features with b) the set of crops and the associated plurality of crop features; and
building, by the one or more hardware processors, the ML crop mapping model for the ROI using training data comprising the labelled plurality of optimal pixels, the corresponding plurality of pixel features and a set of remote sensing images for the ROI, wherein the ML crop mapping model learns the pattern from the training data and the set of remote sensing images to predict associated crop for each input pixel.