University of Twente Student Theses
Agricultural Field Boundary Delineations in Smallholder Farming Systems of Southeast Asia Using Sentinel-2 Data and Convolutional Deep Learning Models
Bhari Shivaprasad, Ashik (2023) Agricultural Field Boundary Delineations in Smallholder Farming Systems of Southeast Asia Using Sentinel-2 Data and Convolutional Deep Learning Models.
|
PDF
4MB |
| Abstract: | Agricultural Field Boundary (AFB) delineation is beneficial for estimating incentives as part of farming schemes. AFB delineation can help formulate innovative micro-agricultural finance programs, agricultural field statistics calculation, crop yield estimation, and other applications of precision agriculture practice(Enclona et al., 2004). Traditional methods used to monitor AFB are time-consuming and labor-intensive since they are based on human field surveys. Furthermore, the diversity of Earth Observation (EO) technology allows for data collection via a wide range of sensors with varied spatial, spectral, and temporal resolutions. Combined with the recent advancements in computer vision and machine learning algorithms, it is convenient to perform the delineation of agricultural field boundaries. Despite the obvious advantages, it is still the abundance of data created by EO sources can cause a variety of problems in processing. Through this research, we create a tailoredworkflowthat efficiently delineates the AFB from pre-processed Sentinel-2EO data built with seasonal statistic-based composites such as geometric median, median, and medoid with the help of CNN (U-Net) and a post-processing method based on graph-based segmentation and contour extraction for polygonization of boundary predictions. |
| Item Type: | Essay (Master) |
| Faculty: | ITC: Faculty of Geo-information Science and Earth Observation |
| Subject: | 38 earth sciences |
| Programme: | Geoinformation Science and Earth Observation MSc (75014) |
| Link to this item: | https://purl.utwente.nl/essays/97164 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page