Advanced Image Processing Methods for Early Detection and Monitoring of Aphid Infestations

The presence of aphids poses a threat to citrus production. Early detection is important to reduce damage and economic losses, which requires the adoption of methods for their detection. In this study, image processing is employed. An algorithm was developed using the LeLePhid database, consisting of 655 images of lemon leaves captured in a field in Ecuador. Implemented in Python and supported by the segmentation tool Removebg, the algorithm employs techniques such as image erosion and the application of color transformations. The results reveal an accuracy of 95% in the set of healthy and infected images. A comparison was made with other pest detection approaches, highlighting the effectiveness of the algorithm in relation to its computational efficiency. These findings justify future research focused on improving image processing or implementing deep learning techniques.