Abstract
In mandarin production, understanding the number of fruit sets is crucial for proper thinning. Thinning helps prevent biennial bearing and maintains the size and quality of the fruit. This study aims to estimate the number of mandarin fruit sets and proposes a system that estimates the fruit set count from RGB videos of mandarin trees. Appropriate frames are extracted from the RGB videos to create mask images of the mandarin fruits. These images are then input into a deep learning model for fruit set estimation to obtain the predicted count of fruit sets.