Font Size: a A A

An Experiment System Of Parking Space Detection Based On Video Analysis

Posted on:2017-01-28Degree:MasterType:Thesis
Country:ChinaCandidate:J Z JiaFull Text:PDF
GTID:2308330485492587Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
Currently, the national life is gradually rich under the impetus of economic.The number of vehicles is also increased year by year, and parking problem is increasingly serious. It has become a research hotspot that how to use modern science and technology to make full use of the parking space resources.We analyze the overseas and domestic research status of intelligence parking lot at first, and then compare some parking space detection methods which are in common use. Video parking space detection technology has obvious advantages compared with other technologies. Aiming at the disadvantages of the traditional parking space detection algorithm, this paper proposed a detection algorithm based on video image processing.This method judge the state of parking spaces using gray features together with texture features. Feature management is the key of parking space detection technology. Extract texture features and gray features from each parking space image, and group them into a union feature vector. Then we employ Principal Component Analysis (PCA) to reduce the dimensionality of the union feature vector. Support Vector Machine (SVM) is employed to train the feature vector and we get a model after this step. Finally, we can use the model to classify video images which need to be detected the status of parking space. Thus, we get information of whether the parking space is occupied, and the parking space information will be displayed in real time. In order to make the model more accord with the actual conditions, the training samples containing video images of different weather, illumination, pedestrians, sundries and shadows.The results show that the algorithm testing accuracy was 98.15%, and can eliminate the bad effects of weather, illumination, pedestrians, sundries and shadows. Furthermore, the proposed algorithm can handle multiple parking spaces at the same time, greatly improving the management efficiency of parking lot. In the end of this paper, we use MATLAB to simulate the parking space detection method, and construct parking space detection experimental system on the basis of simulation.
Keywords/Search Tags:Parking space detection, Gray histogram, Texture histogram, PCA, SVM
PDF Full Text Request
Related items