A Vision based Smart Multihead Weighing Machine for Automated Food Packager System

Using computer vision based automated intelligent systems a set of food processing system can be implemented to reduce sorting and processing time. Earlier times the corps or food materials are weighed manually or mechanically. In this paper the automated weighing and channeling processing has been proposed using vision based system. Images are captured from camera and food materials in each plate has been analysed for weight. A deep learning based segmentation technique has been adapted for segmenting the food materials and then on image analysis the weight of whole material is calculated. UNet model for semantic segmentation has been applied for segmentation approach. After that knapsack algorithm is run on various weights of the materials for proper packaging. The proposed tehcnique has been tried for crops like rice and cereals and compared with original weight, there is over all error margin of 13 % −17 %.