[Cheng-Chung Chang] Rapid Plant Pesticide Detection Technology by Combining Raman Spectroscopy with Deep Learning
Professor Cheng-Chung Chang and his team developed a Raman spectrum preprocessing method that combines wavelet transform denoising with modified polynomial fitting for automatic removal of fluorescence background, thereby improving pesticide identification performance. Integrated with a convolutional neural network (CNN), this approach enabled accurate analysis of both single and mixed pesticides, achieving 99.1% accuracy in mixed pesticide identification and demonstrating strong potential for rapid detection applications. The study was published in Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (Kuo et al., 2025).
Article source: https://www.sciencedirect.com/science/article/abs/pii/S1386142524013283