Raw Oil Palm Fruit Video Dataset in Plantations

This dataset consists of raw video recordings capturing oil palm Fresh Fruit Bunches (FFB) directly from trees in commercial plantations in Indonesia. The dataset provides an unprocessed look at oil palm fruits in various stages of maturity, including Unripe, Underripe, Ripe, Flower, and Abnormal, each representing different phases of the fruit's development, from early growth to full ripeness, with some exhibiting irregular characteristics due to potential disease. Expert validation has confirmed the accuracy of the maturity categorizations. Recorded at 30 frames per second, the videos were captured using a smartphone, with the operator circling the trees to document all visible FFB from multiple angles. The dataset comprises 440 videos, each lasting between 8 seconds and 1 minute and 31 seconds, with a resolution of 320x640 pixels in MP4 format. This ensures the data's compatibility with various digital media platforms. Several factors contribute to the variability in the recorded footage. The height of the oil palm trees varies, leading to different perspectives in the videos. Additionally, the light intensity differs across the videos, with some appearing brighter than others due to variations in sunlight exposure among the trees. Beyond its primary use in assessing fruit maturity, this raw footage has significant potential for further research and technological development. It can serve as valuable training data for machine learning models focused on automated fruit maturity detection, support the advancement of precision agriculture techniques, and contribute to the development of robotic harvesting solutions. This dataset consists of raw video recordings capturing oil palm Fresh Fruit Bunches (FFB) directly from trees in commercial plantations in Indonesia. The dataset provides an unprocessed look at oil palm fruits in various stages of maturity, including Unripe, Underripe, Ripe, Flower, and Abnormal, each representing different phases of the fruit's development, from early growth to full ripeness, with some exhibiting irregular characteristics due to potential disease. Expert validation has confirmed the accuracy of the maturity categorizations. Recorded at 30 frames per second, the videos were captured using a smartphone, with the operator circling the trees to document all visible FFB from multiple angles. The dataset comprises 440 videos, each lasting between 8 seconds and 1 minute and 31 seconds, with a resolution of 320x640 pixels in MP4 form