UAV Remote Sensing for Western Mayhaw Flower Intensity Assessment Presenting Author: Austin Fruge’ Co-Authors: Dr. Cengiz Koparan, Dr. Donald M Johnson, Dr. Amanda McWhirt Abstract. Western Mayhaw (Crataegus opaca) is an emerging economically important fruit in the genus Crataegus due to increased consumption, expanded marketing, and improved cultivars. Further research is needed to expand technology-driven management strategies and investigate its potential as an economical crop for rural and urban landowners in the Southeastern United States. The current methodology for estimating flowering intensity assessment in Western mayhaws is performed with visual observation in the field. However, this methodology is time-consuming, labor-intensive, and subjective. Given the need for a precise methodology for flowering intensity monitoring in Western mayhaws, we developed an open-source image-based phenotyping workflow from Unmanned Aerial Vehicle (UAV) captured images. A subset of Western mayhaw selections were evaluated for blooming intensity during the spring of 2025 in a private orchard near El Dorado, Arkansas. RGB images of Western mayhaw trees during the early flowering stage were collected using a DJI Mavic 3 Enterprise UAV mounted with an RGB digital camera. Each image was processed using an open-source image processing software to estimate the number of flowers. To evaluate the accuracy of this method, the flowering intensity was evaluated through visual flower counting and a visual scale, and compared to image-based flower estimation. Flowering intensity estimated with image segmentation showed a strong correlation with visual flower counting (r= 0.858, p < 0.001), indicating that an increase in visual flower count can be explained with segmented pixel count for any random image. Flower estimation with image segmentation is accurate and provides a standard method, however, it could be time-consuming due to the large image dataset. A semi-automated or fully automated image processing workflow could be developed to increase the efficiency of image processing.
