Application of Artificial Intelligence and PDCA Cycle Method in Quality Control of Food Inspection Process

In a large-scale food production environment, the lack of representativeness of samples and the lack of scientific of random inspections may lead to inspection results that are inconsistent with the actual situation. To this end, this paper improves the inspection efficiency of food inspection process quality control based on artificial intelligence and PDCA (Plan-Do-Check-Act) cycle method. Combining artificial intelligence with PDCA cycle method, this paper proposes a method for improving the efficiency and accuracy of food inspection process quality control. First, the application of data preprocessing and feature engineering in food quality control is studied, especially the key technologies in standardization, normalization and missing value processing. Then, the method of food quality prediction using decision trees and random forests is introduced in detail to predict the eligibility and quality grade of food. Next, this paper explores the application of convolutional neural network (CNN) in deep learning in food appearance inspection, especially in image recognition of discoloration, mold, packaging defects, etc. In the implementation of the method, combined with the four stages of the PDCA cycle, the above artificial intelligence technology is used for real-time monitoring, quality prediction, anomaly detection and process optimization in the planning, execution, inspection and improvement stages respectively. Through experimental application in dairy products, canned foods and fresh vegetables, the results show that the inspection method combining artificial intelligence with the PDCA cycle method significantly improves the efficiency and accuracy of inspection, reduces production costs, and effectively reduces the food defect rate.