Detection of crop fields in Ukraine using S1_MSI index

• Main Author: Vaishnavi, Inbanathan
• Other Authors: Li Fang
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/168326
• Institution: Nanyang Technological University

A remote sensing approach is needed for identifying and classifying crops accurately as the traditional method of experts visiting the field to identify the crops being grown is a labour-intensive and time-consuming process. In recent days, research on cropland detection based vegetative indices are conducted by many researchers, which have great importance as identifying croplands correctly can improve agricultural planning. Empirical studies also show that the temporal characteristics of vegetative indices are one of the main features that help to detect a cropland. Hence, this feature was extracted for the classification. However, unlike the previous studies using publicly available vegetative indices, Yara’s very own S1_MSI index, a more robust index, was used in this thesis to analyse the fields in Ukraine. A classification approach using S1_MSI temporal data, which categorizes pieces of Ukrainian land to either cropland or non-cropland was proposed to detect crop fields accurately. The purpose of it is to ensure the validity of active cropland hectare (area) count. A database containing fields from 5 landuse classes (forest, residential, industrial, quarry, heath) apart from cropland was constructed, and used for cropland features extraction, classification, and accuracy testing. The result of this cropland detection project is quite satisfactory, as the categorization outcomes are quite accurate. However, due to the limited time and author’s knowledge, some other landuse classes that exhibit very close behaviour to a cropland such as orchard are not included, which may reduce the classification accuracy. Hence, much more effort needs to be put in to perform a better categorization approach.