| บทคัดย่อ |
In this modern age, several new methods have been developed, especially in
image processing for agriculture business, which consists of technologies
derived from artificial intelligence (AI) capabilities calledmachine learning.
Classify is a widely used method to analyze patterns, trends, as well as the
body of knowledge from the data visualization. Image classification
application improves discrimination and prediction efficiency. The objective
of this research was to feature extraction of sweet tamarindand compare the
algorithm for classification. This research used images from golden sweet
tamarind species with the use of MATLAB andpython language. The steps
of this research consisted of 1) preprocessing step for finding the distance to
appropriate of the image quality, 2) feature extracting for finding the number
of black pixels and the number of white pixels, perimeter, diameter, and
centroid, and 3) classifying for algorithms' comparison. The results showed
that the camera's distance to the image was 60 cm. The coefficient of
determination was at 0.9956, and thestandarderror of estimatewas
7,424.736 pixels. The conclusion of classification found that the random
forest had the highest accuracy at 92.00%, SD. = 8.06, precision = 90.12,
recall = 92.86, and F1-score = 91.36.
|
|---|