Research of machine learning under different lighting conditions

With the rapid development of information technology, the application of image processing technology is becoming more and more extensive, and color recognition is a very important research direction. Different lighting conditions have great influence on the accuracy of color recognition. The General K-nearest Neighbor (KNN) classification method can calculate a large amount of data but has low classification efficiency. The Bayesian algorithm is an uncertain method, and the classification results are not necessarily accurate. The Otsu algorithm is the best method in statistical analysis and threshold calculation, The Probabilistic Neural Network (PNN) method, which is a kind of supervised network classifier, has fast training speed, more accurate classification, and good fault tolerance. Therefore, this paper combines the Otsu algorithm with PNN classification to obtain optimal color recognition results under different lighting conditions with higher accuracy.