Abstract: In the field of traffic-information acquisition, one pervasive solution is to use wireless sensor networks (WSNs) to realize vehicle classification and counting. By adopting heterogeneous sensors in a WSN, the potential of using complementary physical information to perform more complicated sensing computation can be explored. However, the collaboration among heterogeneous sensors, such as the collaborative sensing mechanism (CSM), is not well studied in current state-of-the-art research. In this paper, the authors design and implement EasiSee, a real-time vehicle classification and counting system based on WSNs. Contributions are as follows. First, the authors propose a CSM, which coordinates the power-hungry camera sensor and the power-efficient magnetic sensors, reducing the overall system energy consumption and maximizing system lifetime. Second, a robust vehicle image-processing algorithm, i.e., a low-cost image processing algorithm (LIPA), is proposed. LIPA reduces environment noise and interference with low computation complexity. In the verification section, the vehicle detection accuracy turned out to be 95.31%, which pave the way for CSM. The time of image processing is around 200 ms, which indicates that the LIPA is computationally economical. With the overall energy consumption reduced, EasiSee achieves classification accuracy of 93%. Based on these experiments and analysis, the authors conclude that EasiSee is a practical and low-cost affordable solution for traffic-information acquisition. Vehicle identification or classification is one of the application areas that come under real time image processing. Vehicle recognition is having the significance in various applications including the traffic monitoring, load monitoring, number plate recognition, vehicle theft prevention, traffic violation detection, management of traffic etc. As the images are captured as primary data source, it can have number of associated impurities which include the background inclusion, object overlapping etc. Because of this, object detection and recognition is always a challenge in real time scenario. In present work, a robust feature based model is presented for feature extraction and classification of vehicle images. The presented model is applied on real time captured image to categorize the vehicle in light, medium and heavy vehicle. Firstly, the vehicle area segmentation is performed and later on the Gaussian filter is applied to extract the image features. This featured dataset is processed under Support Vector Machine (SVM) based distance analysis model for vehicle recognition and vehicle class identification.

Keywords: Support Vector Machine (SVM), low-cost image processing algorithm (LIPA), CSM, WSN.