The quality of palm oil is strongly influenced by the maturity level of the fruit to be processed into palm oil. Many studies have been carried out for detecting and classifying the maturity level of oil palm fruit to improve the quality with the use of computer vision. However, most of these studies use datasets in the form of images of oil palm fresh fruit bunches (FFB) with incomplete categorization according to real conditions in palm oil mills. Therefore, this study introduces a new complete dataset obtained directly from palm oil mills in the form of videos and images with different categories in accordance with the real conditions faced by the grading section of the palm oil mill. The video dataset consists of 45 videos with a single category of FFB videos and 56 videos with a collection of FFB with multiple categories for each video. Videos are collected using a smart phone with a size of 1280 × 720 pixels with .mp4 format. In addition, this dataset has also been annotated and labelled based on the maturity level of oil palm fruit with 6 categories, which are unripe, under-ripe, ripe, overripe, empty bunches and abnormal fruit.

No abstract

The classification of the ripeness degree of oil palm fruit has attracted the attention of numerous researchers. However, there are still many challenges due to constraints in the dataset, methodologies used, and variations in the use of data categories. Detecting oil palm fruit bunches accurately is crucial, given their complex shape and characteristics, particularly when different ripeness categories are present in a pile of oil palm. Most studies utilize oil palm images or the color spectrum of oil palm fruit to classify the level of ripeness. However, these methods are not real-time and lack efficiency. This study proposes a real-time model for determining the ripeness degree of oil palm using a smartphone and video data as input, incorporating modifications to the object detection approach. The research process involves collecting videos of palm oil piles using smartphones in the grading area of the palm oil industry. The videos are then pre-processed and labelled for the object detection and classification process. A detection and classification model is developed using the YOLOv4 approach with several performance improvements, enabling implementation on smartphones. The best-performing model is tested for detecting and classifying the ripeness of fresh fruit bunches using an android-based smartphone. The testing results, based on the mAP value, demonstrate that the YOLOv4 model with 16 quantization performs 12% better than YOLOv4 Tiny. Based on the test results at the grading location, this model can efficiently detect fruit bunches that do not meet the quality standards.INDEX TERMS Hyperparameter tunning, oil palm ripeness, real-time detection, modified YOLOv4.