Detection of Infective Juvenile Stage of Entomopathogenic Nematodes Using Deep Learning

Entomopathogenic nematodes, particularly Steinernema glaseri, are essential for biological pest control, with their infective juvenile (IJ) stage being critical for insect infection. Accurate identification of the IJ stage, typically performed using a stereomicroscope, is time-consuming and prone to human error. This research proposes a deep learning-based method for identifying the IJ stage nematodes from stereomicroscope images. The YOLOv5s architecture is utilized and enhanced by image pre-processing with Contrast Limited Adaptive Histogram Equalization (CLAHE) to improve image quality. Experimental results confirm the effectiveness of our approach, achieving a precision of 0.781, recall of 0.783, and a mean Average Precision (mAP50) of 0.817. These results demonstrate the capability of the deep learning model combined with CLAHE to enhance detection accuracy, reduce processing time, and minimize errors.