Faculty of Engineering, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan.
Global Education Center, Shinshu University, 3-1-1 Asahi, Matsumoto 390-8621, Japan.
Sensors (Basel). 2020 Jan 23;20(3):637. doi: 10.3390/s20030637.
Oil palm ripeness' main evaluation procedure is traditionally accomplished by human vision. However, the dependency on human evaluators to grade the ripeness of oil palm fresh fruit bunches (FFBs) by traditional means could lead to inaccuracy that can cause a reduction in oil palm fruit oil extraction rate (OER). This paper emphasizes the fruit battery method to distinguish oil palm fruit FFB ripeness stages by determining the value of load resistance voltage and its moisture content resolution. In addition, computer vision using a color feature is tested on the same samples to compare the accuracy score using support vector machine (SVM). The accuracy score results of the fruit battery, computer vision, and a combination of both methods' accuracy scores are evaluated and compared. When the ripe and unripe samples were tested for load resistance voltage ranging from 10 Ω to 10 kΩ, three resistance values were shortlisted and tested for moisture content resolution evaluation. A 1 kΩ load resistance showed the best moisture content resolution, and the results were used for accuracy score evaluation comparison with computer vision. From the results obtained, the accuracy scores for the combination method are the highest, followed by the fruit battery and computer vision methods.
油棕成熟度的主要评估程序传统上是通过人工视觉完成的。然而，依赖人类评估者通过传统方法来评估油棕新鲜果实束（FFB）的成熟度可能会导致不准确，从而降低油棕果实的油提取率（OER）。本文强调了水果电池法通过确定负载电阻电压及其水分含量分辨率来区分油棕果实 FFB 成熟度阶段。此外，还对相同的样本进行了基于颜色特征的计算机视觉测试，以使用支持向量机（SVM）比较准确性评分。评估并比较了水果电池、计算机视觉以及两种方法准确性评分组合的准确性评分结果。当测试范围为 10 Ω 至 10 kΩ 的负载电阻电压时，对成熟和未成熟样本进行了测试，从中筛选出三个电阻值并对其水分含量分辨率进行了测试。1 kΩ 的负载电阻表现出最佳的水分含量分辨率，并且使用该结果与计算机视觉进行了准确性评分评估比较。从获得的结果来看，组合方法的准确性评分最高，其次是水果电池和计算机视觉方法。
