In the heart of Peru’s Amazon region, a humble legume known as pajuro, or *Erythrina edulis*, is gaining attention not just for its nutritional value but also for the innovative technology being developed to streamline its harvest. Researchers, led by Hector Pasache from the School of Mechatronics Engineering at Universidad Peruana de Ciencias Aplicadas in Lima, have pioneered an automated system that could revolutionize the sorting of pajuro beans, offering a blueprint for other agricultural products.
The current method of sorting pajuro beans is a labor-intensive process, relying heavily on manual labor to separate seeds based on their maturation stages. This not only slows down production but also introduces variability, which can be particularly problematic in industrial settings. Pasache and his team have addressed this challenge by developing a custom lightweight convolutional neural network (CNN) designed for real-time execution on embedded hardware. This system employs a fixed-region segmentation strategy to prevent overcounting and uses GPIO-based control on a Raspberry Pi 5 to synchronize seed classification with physical sorting in real time.
“The system employs a servo-controlled ejection mechanism to automatically remove defective seeds, ensuring only the highest quality makes it through the sorting process,” Pasache explained. The integrated system combines object detection, image processing, and real-time actuation, achieving a classification accuracy exceeding 99.6% and an average processing time of 12.4 milliseconds per seed.
The implications of this research extend beyond the sorting of pajuro beans. The scalable framework developed by Pasache and his team can be applied to a wide range of agricultural products, offering a solution for color-based grain classification. This technology has the potential to significantly enhance the efficiency and accuracy of sorting processes in the agricultural industry, reducing labor costs and improving product consistency.
“This technology is not just about improving the sorting process for pajuro beans; it’s about creating a scalable model that can be adapted to various agricultural products,” Pasache added. The research, published in the journal *AgriEngineering* (translated to English as Agricultural Engineering), highlights the potential for computer vision and convolutional neural networks to drive innovation in the agricultural sector.
As the demand for high-quality agricultural products continues to grow, the need for efficient and accurate sorting technologies becomes increasingly important. The work of Pasache and his team represents a significant step forward in this field, offering a glimpse into the future of agricultural automation. The commercial impacts for the energy sector, particularly in terms of reducing labor costs and improving product consistency, are substantial. This research not only shapes the future of pajuro bean sorting but also sets a precedent for the broader application of computer vision and CNN technologies in agriculture.