Research On The Detection Of Parking Lots In Remote Sensing Images Based On Deep Learning

With the continuous development of economy and the continuous growth of private car ownership,the contradiction between the layout of traffic network and the adaptability of urban planning is highlighted.The rapid growth of traffic demand far exceeds the followup construction speed of infrastructure in the field of urban traffic.The increasingly severe problem of "difficult parking" is a major obstacle for people to travel.At the same time,the rapid development of satellite remote sensing technology has become the main technical means and data source of space to ground observation.With the help of remote sensing data,corresponding detection can be completed according to different tasks.In order to alleviate the problem of "parking difficulty",it is a new challenge for deep learning to find parking area and parking space resources.Therefore,in this paper,the deep learning theory is applied to the recognition of parking lot in remote sensing image,and the vehicle detection method based on the deep learning is given.At the same time,the parking area and parking space resources under the remote sensing data are discovered and recognized with the help of data mining algorithm.Aiming at the shortcomings of low accuracy and slow speed of traditional target detection algorithm in recognition of small targets in remote sensing image,this paper combines the depth learning theory with remote sensing image target detection in any way,uses the Darknet depth learning framework to build and improve the regression based Yolo model,adds the residual structure,makes the model map reach 94%,adjusts the model input limit,and adds the figure At the same time,the remote sensing data set for experiment is established,and the vehicle detection model of remote sensing image is obtained.According to the output results of remote sensing image vehicle detection model,data mining technology,remote sensing image parking area discovery method and remote sensing image parking space prediction method are used to model and analyze the result set,and finally complete the target detection task under remote sensing image.To solve the problem of parking lot selection in remote sensing image,this paper designs remote sensing image vehicle detection,remote sensing image parking area discovery,remote sensing image parking space quantity prediction,to solve the task of parking area and parking space discovery under remote sensing image.