The increasing demand for high-quality agricultural produce necessitates the modernization of sorting processes, traditionally reliant on manual labour. This study presents the development of a smart fruit and vegetable sorting machine utilizing advanced machine learning and computer vision technologies to enhance sorting accuracy, throughput, and efficiency. The methodology encompasses a systematic approach, including the design and configuration of a conveyor system, implementation of imaging sensors, and the integration of a convolutional neural network for real-time classification of produce. A dataset of 10,000 labelled images was utilized to train the model, which achieved an impressive sorting accuracy of 95% and a throughput of 120 items per minute during testing. The machine demonstrated a low error rate of 5%, underscoring its effectiveness in minimizing post-harvest losses and ensuring quality control. These results highlight the significant advantages of automation in agricultural practices, surpassing traditional manual sorting methods in both speed and reliability. Additionally, an economic feasibility analysis indicated the potential for substantial cost savings in labour and reduced spoilage, making the technology viable for small and medium-sized farms. The findings of this research demonstrate that the smart sorting machine is a transformative solution for contemporary agriculture, addressing critical challenges in sorting efficiency and accuracy. Future work is recommended to explore advanced imaging techniques, real-time monitoring systems, and broader applications across diverse crop types. By embracing these innovations, the agricultural sector can enhance productivity, sustainability, and overall profitability, ultimately contributing to a more efficient food supply chain.
Read full abstract