View Item 
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        JavaScript is disabled for your browser. Some features of this site may not work without it.

        Browse

        All of UU Student Theses RepositoryBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

        Crop Monitoring in Intercropping Systems: Assessing Maize Height patterns using UAV-imagery

        Thumbnail
        View/Open
        Thesis_LaraGillham.pdf (3.923Mb)
        Publication date
        2024
        Author
        Gillham, Lara
        Metadata
        Show full item record
        Summary
        In the quest for sustainable agriculture, intercropping is emerging as a promising solution to monoculture. Nevertheless, the existing research on phenotyping in intercropping systems is insufficient. Unmanned Aerial Vehicles (UAVs), especially when equipped with RGB and LiDAR imagery, offer a promising avenue for estimating plant height. The accuracy of these two methods in estimating maize height in a strip intercropping field, focusing on individual and row levels, and observing height patterns throughout the growing season were estimated. A crop height model (CHM) for both types of imagery was created by subtracting a digital surface model (DSM) from a digital terrain model (DTM). Various percentiles and buffer sizes were tested, and the correlation between UAV-estimated plant height and ground truth plant height was evaluated using the coefficient of determination (R²) and root mean square error (RMSE) to determine the most suitable parameters. The results indicated that LiDAR was more accurate in capturing maize tassel height, with higher results for individual crop observations (R²= 0.9, RMSE= 13.41cm) than for row observations (R² = 0.89, RMSE = 20.89cm). RGB imagery yielded an R² of 0.88 and RMSE of 27.88 cm for individual crop height and an R² of 0.91 and RMSE of 25.97cm for row observations. Once the optimal parameters are identified, the height patterns in rows and over time could be observed within the field experiment. Height variations were noted within the experimental field, with southern rows typically having lower height values. The influence of neighboring crops was also apparent, as maize plots exhibited lower height values with beans as a neighboring crop compared to neighboring maize. To increase accuracy and automation, future studies should consider using tassel detection. Growth curves were observed as highly unique to the different genotypes. This study suggests that LiDAR imagery can offer a reliable assessment of individual maize heights in intercropping fields and that the noticeable height patterns between rows should be taken into account when discerning differences in the behavior of different crops and cultivars in intercropping systems.
        URI
        https://studenttheses.uu.nl/handle/20.500.12932/46414
        Collections
        • Theses
        Utrecht university logo