Coin Image Recognition Based on K-Nearest Neighbor Algorithm

Inhalt:
At present, the coin recognition system has problems, such as high hardware cost and complex system structure. To solve these problems, this paper proposes a coin image recognition system based on K-nearest neighbor algorithm, the system has the advantages of low cost and high efficiency. Firstly, collecting the images of 1 angle coin, 5 angle coin and 1-yuan coin, then use the HSV color space model to obtain the color space histograms of the 1 angle coin, 5 angle coin and 1- yuan coin, and distinguish the 5 angle coin according to the H and S components of the color space histogram. Different perimeters are obtained by calculating the closed edges of the images of the 1 angle coin and the 1-yuan coin to distinguish the 1 angle coin and 1-yuan coin, and then use the K-nearest neighbor algorithm to classify and train the feature data of the coin, and finally randomly select 30 untrained images of each coin for the experiment. The experimental result shows that the system can effectively realize coin image recognition.