UAV Images for Agriculture Land Parcel Delineation through Edge Detection Algorithm: A Case Study of Hilly and Terai Regions

The use of unmanned aerial vehicles (UAVs) for remote sensing applications has gained significant attention in recent years. One important aspect of UAV-based remote sensingis the accurate delineation of parcels, which is essential for a wide range of applications, including land use planning, agricultural monitoring, and cadastral map preparation among several others. The on-screen manual digitization method can delineate parcelswith great precision. However, the method is labor-intensive, time-consuming, and expensive. Alternatively, it can be achieved by utilizing automated algorithms. In this study, an edge detection algorithm is employed to delineate agriculture parcels using UAV images in the ENVI platform. The algorithm uses a pre-programmed algorithm to automatically detect and delineate field boundaries. The delineated boundaries were cleaned and refined by smoothing the polygon and the line vectors. The obtained parcel boundaries and their geometric parameters were assessed against manually digitized parcel boundaries using the same UAV ortho-mosaic. The method was tested on two scenarios: i) Terai farms having flat topography and small dikes and ii) Hill farms having undulated terrain in a terraced farming structure and crowded parcels. The mean of the percentage change in area for the land parcel was found to be 2.43% and 4.69% respectively for Terai and Hilly regions. Similarly, the mean of the percentage change in the perimeter of the land parcels were 8.82% and 2.43% respectively for Terai and Hilly regions. The study demonstrated the feasibility of using UAV images for agriculture land parcel delineation and highlighted the potential of an in-built edge detection algorithm as a time-efficient and reliable alternative to manual digitization if refinement and proper selection of algorithm and parameters are done. Thus, automated algorithms can be utilized to reasonably delineate agriculture parcels from UAV images.