Apricot Disease Detection with Convolutional Neural Network

Apricot is a stone fruit grown in temperate climates and possesses high economic value globally. However, diseases and pests pose substantial threats to apricot production, undermining both crop quality and overall yield. As these pressures intensify, they further compromise fruit development and reduce harvest quantities, negatively affecting market value and productivity. In particular, canker, coryneum beijerinckii, drying symptom, and monilinia laxa stand out as the four main diseases that markedly reduce quality and yield worldwide. Therefore, early diagnosis and targeted management strategies for these diseases are critically important for preventing epidemic spread and ensuring efficient resource utilization. In this study, a novel deep learning-based convolutional neural network model is proposed for the detection of diseased apricot images. The proposed CNN model was tested on a publicly available dataset, meticulously compiled under real field conditions and encompassing the aforementioned four apricot diseases. The proposed model achieved a high accuracy rate of 97.74% in the detection and classification of diseases. It provided 8.1% to 21.16% higher accuracy than traditional image processing-based approaches in the literature. Furthermore, the final model achieved 0.44% to 23.87% higher performance compared to some CNN models. These results indicate that the proposed CNN model can provide rapid and reliable decision support in disease detection.