• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.12
• /
• pp.5990-6008
• /
• 2019

Recognizing license plate from its traffic camera captured images is one of the most important aspects in many traffic management systems. Despite many sophisticated license plate recognition related algorithms available online, license plate recognition is still a hot research issue because license plates in each country all round the world lack of uniform format and their camera captured images are often affected by multiple adverse factors, such as low resolution, poor illumination effects, installation problem etc. A novel method is proposed in this paper to solve the inclination problem of license plates in their camera captured images through four parts: Firstly, special edge pixels of license plate are chosen to represent main information of license plates. Secondly, least square methods are used to compute the inclined angle of license plates. Then, coordinate rotation methods are used to rotate the license plate. At last, bilinear interpolation methods are used to improve the performance of license plate rotation. Several experimental results demonstrated that our proposed method can solve the inclination problem about license plate in visual aspect and can improve the recognition rate when used as the image preprocessing method.

• Journal of Korean Society of Transportation
• /
• v.27 no.5
• /
• pp.209-221
• /
• 2009

Travel time estimation under given traffic conditions is important for providing drivers with travel time prediction information. But the present expressway travel time estimation process cannot calculate a reliable travel time. The objective of this study is to estimate the path travel time spent in a through lane between origin tollgates and destination tollgates on an expressway as a prerequisite result to offer reliable prediction information. Useful and abundant toll collection system (TCS) data were used. When estimating the path travel time, the path travel time is estimated combining the link travel time obtained through a preprocessing process. In the case of a lack of TCS data, the TCS travel time for previous intervals is referenced using the linear interpolation method after analyzing the increase pattern for the travel time. When the TCS data are absent over a long-term period, the dynamic travel time using the VDS time space diagram is estimated. The travel time estimated by the model proposed can be validated statistically when compared to the travel time obtained from vehicles traveling the path directly. The results show that the proposed model can be utilized for estimating a reliable travel time for a long-distance path in which there are a variaty of travel times from the same departure time, the intervals are large and the change in the representative travel time is irregular for a short period.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.163-177
• /
• 2019

As smartphones are getting widely used, human activity recognition (HAR) tasks for recognizing personal activities of smartphone users with multimodal data have been actively studied recently. The research area is expanding from the recognition of the simple body movement of an individual user to the recognition of low-level behavior and high-level behavior. However, HAR tasks for recognizing interaction behavior with other people, such as whether the user is accompanying or communicating with someone else, have gotten less attention so far. And previous research for recognizing interaction behavior has usually depended on audio, Bluetooth, and Wi-Fi sensors, which are vulnerable to privacy issues and require much time to collect enough data. Whereas physical sensors including accelerometer, magnetic field and gyroscope sensors are less vulnerable to privacy issues and can collect a large amount of data within a short time. In this paper, a method for detecting accompanying status based on deep learning model by only using multimodal physical sensor data, such as an accelerometer, magnetic field and gyroscope, was proposed. The accompanying status was defined as a redefinition of a part of the user interaction behavior, including whether the user is accompanying with an acquaintance at a close distance and the user is actively communicating with the acquaintance. A framework based on convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent networks for classifying accompanying and conversation was proposed. First, a data preprocessing method which consists of time synchronization of multimodal data from different physical sensors, data normalization and sequence data generation was introduced. We applied the nearest interpolation to synchronize the time of collected data from different sensors. Normalization was performed for each x, y, z axis value of the sensor data, and the sequence data was generated according to the sliding window method. Then, the sequence data became the input for CNN, where feature maps representing local dependencies of the original sequence are extracted. The CNN consisted of 3 convolutional layers and did not have a pooling layer to maintain the temporal information of the sequence data. Next, LSTM recurrent networks received the feature maps, learned long-term dependencies from them and extracted features. The LSTM recurrent networks consisted of two layers, each with 128 cells. Finally, the extracted features were used for classification by softmax classifier. The loss function of the model was cross entropy function and the weights of the model were randomly initialized on a normal distribution with an average of 0 and a standard deviation of 0.1. The model was trained using adaptive moment estimation (ADAM) optimization algorithm and the mini batch size was set to 128. We applied dropout to input values of the LSTM recurrent networks to prevent overfitting. The initial learning rate was set to 0.001, and it decreased exponentially by 0.99 at the end of each epoch training. An Android smartphone application was developed and released to collect data. We collected smartphone data for a total of 18 subjects. Using the data, the model classified accompanying and conversation by 98.74% and 98.83% accuracy each. Both the F1 score and accuracy of the model were higher than the F1 score and accuracy of the majority vote classifier, support vector machine, and deep recurrent neural network. In the future research, we will focus on more rigorous multimodal sensor data synchronization methods that minimize the time stamp differences. In addition, we will further study transfer learning method that enables transfer of trained models tailored to the training data to the evaluation data that follows a different distribution. It is expected that a model capable of exhibiting robust recognition performance against changes in data that is not considered in the model learning stage will be obtained.