Browsing by Author "Baculo, Maria Jeseca C."

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Automatic mango detection using Image processing and HOG-SVM
    (Association for Computing Machinery, 2018-12) Baculo, Maria Jeseca C.; Marcos, Nelson
    Mango is an agricultural produce with high export value as it is being consumed internationally. To ensure its production yield, the manual handling and classification tasks should be performed with precision and care by local farmers. Image processing and machine learning has improved the way classification, defect detection, and yield approximation are handled. Detection is considered as an initial step prior to performing these tasks. This paper presents an automatic mango detector by combining a Support Vector Machine (SVM) classifier trained with Histogram of Oriented Gradients (HOG) features and image segmentation. The image segmentation performed on both HSV and RGB color spaces using image processing techniques achieved a mean IoU of 0.7938. A HOG-SVM based classifier was trained and achieved an F-score of 89.38%. Results show that combining segmentation with HOG-SVM can detect and localize healthy and defective mango images with different background color and illumination.
  • Thumbnail Image
    Item
    Image-based classification and segmentation of healthy and defective mangoes
    (SPIE - The International Society for Optical Engineering, 2019) Baculo, Maria Jeseca C.; Ruiz, Conrado
    The use of image processing and classification for agricultural applications has been widely studied and has led to work such as the automatic grading of fruit and vegetables, yield approximation and defect detection. Image segmentation is one of the first steps to identify the region of interest within an image. This paper presents an approach to automatic segmentation and classification of healthy and defective Carabao mangoes. K-means, range filtering and color-channel segmentation were utilized so that the varying texture and color of mangoes due to the surface defects can be considered. Results show that the proposed technique performs better than the classical K-means segmentation. The performance of segmentation step has a considerable influence on the precision of the classification model. Segmented and not segmented images were trained using KNN, SVM, MLP and CNN. The experiments showed that the models performed better when trained with segmented images.
  • Thumbnail Image
    Item
    Image-based Mangifera indica pathogen recognition using artificial intelligence
    (Association for Computing Machinery (ACM), 26-08-2024) Baculo, Maria Jeseca C.; Rivera, Nema Rose D.; Cuison, Floribeth P.
    Mangifera Indica, holds significant global export value. This study focuses on implementing object detection frameworks to identify five surface defects in this mango variety, which is crucial for maintaining its export quality. The methodology involves training four object detection frameworks. Results show that the modified region extraction technique, which uses adaptive binarization and morphological operations catered to detect mango surface defects, with EfficientNet as the base learner, demonstrated improved accuracy with a mean Average Precision (mAP) of 0.842 at an Intersection over the Union (IoU) threshold of 0.75.