A framework for outdoor parking detection using hybrid model and its implementation

Abstract

In this era of modernization, there is urgent need of automation in the field of detecting vacant lots in the outdoor parking area. Therefore, an optimized Parking Management System (PMS) is required which is based on real-time image processing techniques. Hence, in this research, a framework for outdoor parking detection using hybrid model is designed. In this research work, we used publically available dataset quotPKLotquot. Various image enhancement, image segmentation and feature extraction techniques present in literature are discussed and applied. Further, various classifiers like ANN, KNN, NBC and SVM available in the literature are elaborated and applied for classification on a few thousand samples, selected randomly from the dataset. A comparison of various machine learning techniques based on various parameters is also presented and experimental results are discussed. The experimental results shows that SVM provides the best results. Thereafter, a Hybrid Classification Model based on the results of machine learning techniques is designed which improves the performance of the system at the cost of increased training and testing time. Finally, we utilized the deep learning technique, Convolutional Neural Network (CNN) for the classification problem. A new approach quotSalNet+CNNquot is proposed in which CNN is directly applied on the Region of Interest extracted by saliency detection technique. The experimental results, compared with other existing models, show that the proposed method is significant and provides optimized results over the existing state-of-art techniques. newline