Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 49278 (05)
- University: Sharif University of Technology
- Department: Electrical Engineering
- Advisor(s): Gholampour, Iman
- Abstract:
- In recent years, Intelligent Transportation Systems (ITS) have received special attentions both in research and in commercial areas. Increased infrastructure facilities, like surveillance cameras, has made this concept even more attainable than before. In this respect, the ability to automatically extract information from traffic images, as one of the key inputs of ITSs, is of great importance. With an increased number of surveillance cameras and the need for more accurate information regarding the road users and their interactions, in order to better city traffic management, building and repairing roads, trip time estimation, number of people per roads estimation and etc, using human operators, while not resulting in acceptable outputs, imposes unnecessary high costs. Regarding these issues, in this research a machine vision based automatic system for vehicle detection and classification has been proposed. The input images, coming from Tehran surveillance cameras, are of low quality and resolution, which makes the problem even more challenging. Most of the variant methods in vehicle detection literature are more suitable for high resolution inputs. In these methods, the detection process is based on edge-like features which attain poor results on low resolution images. Therefore, in this research, a deep learning based method is proposed for vehicle detection and classification. In recent years, deep learning, due to its ability to extract complex features from the inputs, has achieved state of the art results in multiple areas such as, image, natural language and sound processing. In the proposed method, the detection and classification process is done in one step and using a single deep neural network. This unified approach allows easier generalization and optimization of the method compared to the common three step approach in literature, hypothesis generation, feature extraction and classification. Also, as is shown in the thesis body, a single model has been trained for two different views, showing that the method is view-independent, achieving a precision and recall of higher that 88 and 93 percents, in detecting the vehicles and classifying them into three types of bus, truck and sedan, on both views
- Keywords:
- Intelligent Transportation System (ITS) ; Surveillance Cameras ; Machine Vision ; Deep Learning ; Vehicle Classification
Digital Object List
-
محتواي کتاب
- view
Bookmark