Optimizing the Object using Real-Time Computer Vision and Neural Network

Poultry and food processing manufacturing units have several automated and monitoring processes of the food that go into packaging. However, at the packaging section, manual human intervention is needed to ensure that the correct amount of food, in terms of count and weight, is placed into every package. This is not without error. Hence, an accurate real-time measurement of sample object counts and weight is critical for optimizing processing efficiency and automating workflows in the production chain. The system identifies the food object, counts it, and weighs it before packaging the batch. In this work, we present a novel approach that integrates an Ultralytics You Only Look Once (YOLO) v10 model, a convolutional neural network (CNN)-based object detection framework, with an automated weighing system and dashboard to optimize quality control.