Smart City Image Recognition Method Using Convolutional Neural Networks

In the new period, with the continuous development of society, urban management should also keep pace with the times, take the initiative to learn new ideas, introduce new technologies, and innovate and optimize in the management mode and management content in order to better adapt to the needs of modern urban construction. To this end, the competent authorities must pay close attention to the latest technology, in particular the use of “5G+Big Data”, in order to achieve the ultimate goal of automated and intelligent management by means of a variety of information technology tools, thereby promoting the modernization of cities. Based on this, this paper conducts research on the application of 5G+big data in smart city management, which has important theoretical and practical significance. In smart city case recognition, the object detection method is able to accurately determine the category of an object based on its different characteristics through repeated learning, which is important in common cases such as illegal parking of vehicles and illegal posting of advertisements. In this paper, a novel object detection method based on Convolutional Neural Network (CNN) is presented and evaluated. The improved U-Net (Convolutional Networks for Biomedical Image Segmentation) network performs excellently in semantic segmentation tasks, with an average accuracy of 91.4%, while the CLRNet-L (Cross layer Refined Network with Large Kernel Attention for Object Detection) network performs well in object detection, with an average accuracy of 93.5% and processing speed that meets real-time monitoring requirements. The experimental results show that the model introduced in this paper achieves efficient and accurate detection and classification in smart city image recognition, which has a wide application potential.