Improved AOA Algorithm to Optimize Image Entropy for Image Recognition Model

Abstract

With the continuous development of computer vision, the application of image recognition technology is becoming increasingly widespread. An edge detection image recognition model based on improved artificial bee colony algorithm has been proposed. Firstly, the identification process of artificial bee colonies is designed. To solve the algorithm easily falling into local optima, a GA with a global search strategy is further improved, achieving an improvement in model operation speed and coherence. Moreover, the target detection and localization methods are selected. The Canny operator and line fitting method are ultimately determined for image search and localization. To further verify the reliability of the improved artificial bee colony algorithm, simulation experiments are conducted on the MATLAB platform. The experimental results show that under 0.1 noise, the improved artificial bee colony algorithm has better recognition accuracy, compared to the particle swarm algorithm. The calculation time is reduced by 7.35s. In summary, the improved artificial bee colony algorithm has the best recognition accuracy and noise resistance performance.