<![CDATA[This study aims to enhance the quality of straw mushroom images by applying a Hybrid White Balance (HWB) preprocessing method to improve classification accuracy. Variations in illumination often introduce color distortion, which negatively affects feature extraction and reduces the performance of machine learning models. Therefore, robust preprocessing techniques are required to handle lighting inconsistencies and improve image quality. In this study, the HWB method is combined with normalization and histogram equalization to produce more consistent visual representations. Straw mushroom images were collected under varying lighting conditions from different agricultural environments. The preprocessing stage includes color correction using HWB followed by normalization to reduce variability. The processed images were then classified using a Convolutional Neural Network (CNN). The results show that preprocessing significantly affects classification performance. The model without preprocessing achieved an mAP@0.5 of approximately 0.927, while Standard White Balance improved performance to around 0.978 in terms of precision, recall, and F1-score. The best results were obtained using HWB, achieving precision of approximately 0.996, mAP of about 0.994, and F1-score around 0.9395, indicating more accurate and robust classification. Additionally, image quality evaluation shows that HWB reduces Mean Squared Error (MSE) and increases Peak Signal-to-Noise Ratio (PSNR) and Signal-to-Noise Ratio (SNR), outperforming standard preprocessing methods. In conclusion, HWB-based preprocessing effectively enhances both image quality and classification performance. This method has strong potential as a reliable preprocessing approach for agricultural image analysis, particularly under varying illumination conditions, and can support the development of more robust and adaptive classification systems.]]>
Copyrights © 2026
