Detection and Counting of Olive Trees in Images Taken by an Unmanned Aerial Vehicle

Unmanned aerial vehicles are speedily improving and taking part in remote sensing applications. Precision agriculture applications aim to increase yield and profitability without wasting limited resources and challenging climate change. Olive tree detection and counting is a precision agriculture application. It helps estimate production yield. In this study, the olive tree images were taken with a camera mounted on an unmanned aerial vehicle. Image analysis techniques were implemented to extract and specify the olive tree crown in a few stages. Firstly, uneven illumination was corrected to get uniform light distribution in the image. The next standard deviation filter was utilized to obtain the olive tree boundaries. Following morphological filling to remove discontinuities in the image, the circular Hough transform was employed to detect circular objects. Finally, the detected circular objects were counted automatically. This approach is straightforward, fast, and effortlessly carried out with direct steps. It can beat a few difficulties, like unbalanced light distribution due to natural conditions, time consumption, and many overlapped trees. This technique shows high detection and counting rate for the delineation, detection, and counting of olive trees. The true-positive rate and precision were 94% and 95% on average for test images.