Smart pesticide dosing system for cocoa plant based on computer vision and deep learning technologies.
The cocoa plant is susceptible to various diseases caused by bacteria, viruses and pests, resulting in yield losses. Visual analysis is a subjective and time-consuming process. In addition, farmers use improper pesticides to prevent diseases, causing degradation of the plant and soil quality. To overcome these problems, a recent study by a group of Indian researchers (Arakeri et al., 2024) aimed to develop an intelligent system, based on computer vision and deep learning techniques (AI), for the automatic detection of plant diseases and the subsequent dosage of the most appropriate pesticides.
The proposed system has proven to achieve a very high accuracy, with a maximum value (above 97%) obtained by applying the EfficientNetB1 model. According to the authors, the proposed system can assist farmers in the timely detection of plant diseases, preventing yield losses and reducing the use of inappropriate pesticides. Therefore, this system can promote sustainable agriculture practices and contribute to the long-term viability of the cocoa farming community. Furthermore, this tool can be enhanced by providing weather information of the cropland for pesticide application and classification of diseases under the pathogens category.
Development of a computer vision system for categorizing of cocoa fruits
The purpose of a recent study, carried out by a group of Colombian researchers (Rodriguez-Umaña et al., 2024), was to develop and implement an intelligent prototype for the classification of cocoa fruits based on their size and ripeness. The system uses a 1 HP motor to transport the fruits from a loading hopper to the classification area, where a Full HD camera captures detailed images from various angles.
These images are analysed by specialized software assisted by a LED light system for brightness control. According to the results, by implementing this system it was possible to significantly reduce classification times: For a sample of 900 kg, manual classification took approximately 6 hours, while the with the automated system the same process was completed in 1 hour and 45 minutes. Similarly, the prototype also reduces classification errors and, therefore, product losses.
According to the authors, digital image processing and the implementation of digital filters have achieved minimal errors in the calibration of fruit and classification by ripeness compared to other conventional methods. This simple automation also enhances working conditions by reducing manual intervention, optimizing processes and lowering operational costs. By implementing this system, companies can meet international quality standards, facilitating exports and improving their global competitiveness.
References: Arakeri et al., Environmental Research Communications, 6, 2024, 075003; L.A. Rodriguez-Umaña et al., Vision Electronica, 18, 2024, 1-18.
