A method for smoke detection in video is proposed. The camera monitoring the scene is assumed to be stationary. With the atmospheric scattering model, dissipation function is reflected the transmissivity between the background objects in the scene and the camera. Dark channel prior and fast bilateral filter are used for estimating dissipation function which is only the function of the depth of field. Based on dissipation function, visual background extractor (ViBe) can be used for detecting smoke as a result of smoke’s motion characteristics as well as detecting other moving targets. Later on, various characteristics of smoke are extracted. Color feature and high-frequency energy based on wavelet transformation is selected to constitute the final recognition vectors, and Support vector machine (SVM) is used as a classification model. The final experimental results show that the accuracy rate of this method for smoke detection can reach 90.1%.
Vehicle color recognition is easily affected by subtle environmental changes. The existing recognition methods cannot achieve an accurate result. A high-accuracy vehicle color recognition method using a hierarchical fine-tuning strategy for urban surveillance videos is proposed. Different from the conventional convolutional neural networks-based methods, which usually obtain a single classification model, the proposed method combines pretraining and hierarchical fine-tunings to obtain different classification models that can adapt to the change of illumination conditions. First, the GoogLeNet is pretrained using the ILSVRC-2012 dataset to obtain the initial weight parameters of the network. During the first stage of fine-tuning, the whole vehicle color dataset is used to fine-tune the pretrained results to get the initial classification model. Then, an image quality assessment method is proposed to evaluate the illumination conditions of the image. The whole vehicle color dataset is divided into some subdatasets according to the evaluation results. The second stage of fine-tuning is performed on the initial classification model using each subdataset. Thus, the final classification models for the subdatasets are obtained. The experimental results on different databases demonstrate that the recognition accuracy of the proposed method can achieve superior performance over the state-of-the-art methods.
Thermal-response amorphous polymer is a typical soft active material where the glass transition occurs as temperature
changes. In the process of glass transition, the amorphous polymer will go through the transitions between the glassy
state and the rubber state. Furthermore, as subjected to a combining load of the mechanics and temperature, the
materials will experience deformation and glass transition simultaneously and exhibit some peculiar macroscopic
response. To describe this behavior, we consider the evolution of natural configurations as the main mechanism for
such thermo-mechanical behavior, and develop a thermo-mechanical constitutive model of polymer under mechanical
and thermal loads. As application the proposed model, we investigate the uniaxial thermo-mechanical behavior of the
soft active polymers. It is shown that the model is effective for describing the coupling behavior of material evolution
with large deformations in the glass transition region.