|
Classification of Guanxi Mandarin Orange Grades using Machine Vision Algorithms |
|---|---|
| รหัสดีโอไอ | |
| Creator | Fulian Huang |
| Title | Classification of Guanxi Mandarin Orange Grades using Machine Vision Algorithms |
| Contributor | Jialin Xie, Shijun Jie, Nattawoot Suwannata |
| Publisher | Faculty of Informatics, Mahasarakham University |
| Publication Year | 2568 |
| Journal Title | Journal of Applied Informatics and Technology |
| Journal Vol. | 7 |
| Journal No. | 2 |
| Page no. | 305-324 |
| Keyword | Classification, Defect, Detection, Flaw, Machine Vision System, Mandarin Orange, Skin |
| URL Website | https://ph01.tci-thaijo.org/index.php/jait |
| Website title | Journal of Applied Informatics and Technology |
| ISSN | 3088-1803 |
| Abstract | This article proposes a method for categorizing Mandarin orange grades based on Chinese standards using a computer vision system that integrates both hardware and software components. A mechanical roller-flipping device adjusts the Mandarin orange’s position in various orientations. Subsequently, a machine vision system acquires thirty photographs of mandarin orange skin from various viewpoints and employs many processing approaches, such as image acquisition, blob analysis, preprocessing, segmentation, and feature extraction. The process of classifying oranges involves applying techniques such as morphology, median filtering, and the Fourier transform to identify and analyze pixels that represent imperfections on the surface of the orange. Then the faulty pixels are transformed into the diameter and the area of the faults in order to classify them for grading. The experiment demonstrates that the diameter and rectangular regions can be utilized to categorize Mandarin oranges into three grades: Special Grade, Grade 1, and Grade 2. Grade 3 can be determined by measurement of the diameter and calculation of the percentage of the faulty region in the orange peel. The overall recognition accuracy by the system is 87.5%. This experimental method can accurately identify defects in the skin of oranges, reducing labor costs and the error rate of manual identification for enterprises. |
