Applied Mechanics and Materials Vols. 701-702

Abstract: As the change detection based on Synthetic Aperture Radar (SAR) images that are difficult and very limited to acquire labeled samples are of low detection rate and high error rate, Thus a progressive transductive SVM algorithm based on original feature space for unsupervised change detection of SAR images is proposed. The pseudo-training set of the difference image is obtained using K-means clustering method without any prior information; Starting from these initial seeds, the progressive transductive SVM performs change detection in the original multitemporal feature space by gradually considering unlabeled patterns in the definition of the decision boundary between changed and unchanged pixels according to a transductive inference algorithm. Using dynamic region labeling rule, the algorithm not only achieves its rules of progressive labeling and dynamic adjusting, but also raises its speed at the same time. Experimental results obtained on different multitemporal SAR images show that, transductive inference algorithm that extract the information of unlabeled patterns improve the SVM classifier accuracy. These results confirm the effectiveness of the proposed approach.

Abstract: This paper proposes a method of global phase unwrapping used in multi-frequency three-dimensional (3D) measurements. In this method, three kinds of optical information which change in accordance with trigonometric function (sine or cosine) to the objects. The optical information cycles is P1, P2 and P3. Each waveform should 4-8 steps phase shifts. Then, calculate the phase value of each cycle. The composited phase value of two cycles and the final composited phase value of three cycles are calculated by image shift of each cycle. Finally, calculate the global phase value of each cycle based on the composited phase, thus, all the 3D coordinates of objects can be obtain after 3D reconstruction. The proposed method can solves the object surfaces color changing largely in 3D measurements. This method realizes high precision measurement without spray developer and achieves the protection of the environment.

Abstract: In the field of chemical testing, it is of great importance to improve the accuracy and efficiency and reduce the risk caused by artificial factor with modern intelligent methods, which are critical to the standard development of testing labs. In this paper, intelligent control during the whole process of chromatography and mass spectrometry analysis with cloud computing technology was realized and discussed in detail. The intelligent testing system could be applied to the chemical analysis labs and spread to other fields.

Abstract: With the improvement of living standards, people are more concerned about the quality of edible oil. Compared with ordinary edible oil, recycled cooking oil contains heavy metal elements and other impurities. It means that recycled cooking oil’s conductivity is much higher than ordinary cooking oil. According to this feature, system of heavy metals detection and quality analysis based on conductivity of recycled cooking oil have been developed. Taking MCU as the core, this system has achieved functions of conductivity information acquisition, signal processing and human-computer interaction. Adopting bipolar pulse as excitation power supply, the system is aimed at weakening the capacitance effect and polarization effect effectively. Temperature compensation circuit is designed in order to reduce the influence of temperature change on the conductivity measurement. The experimental results show that the detection system has good effects on testing recycled oil which contains different heavy metal components. This detection system has a certain application and promotion value with advantages of low cost, high reliability, high accuracy.

Abstract: With the high humidity caused by the urban microclimate, the visibility is greatly reduced due to the fog weather as well as the hazy weather formed by the accumulation of pollutant and vapor. In order to evaluate how the fog and haze weather affects outside lighting, a hazy simulation environment is introduced in this paper. An ultrasonic humidifier is used to produce droplets for fog simulation. Haze is simulated by using air blower to blow plant ashes. Transmittance of several different light sources under the simulated fog and haze conditions is shown. The results can be used to study outside lighting.

Abstract: In this paper, subspace identification methods are proposed to estimate the linear tire cornering stiffness, which are only based on the road tests data without any prior knowledge. This kind of data-driven method has strong robustness. In order to validate the feasibility and effectiveness of the algorithms, a series of standard road tests are carried out. Comparing with different subspace algorithms used in road tests, it can be concluded that the front tire cornering stiffness can be estimated accurately by the N4SID and CCA methods when the double lane change test data are taken into analysis.

Abstract: This paper proposes a method of realizing a moving vehicle video detecting system based on DSP processor, and describes the system's hardware architecture and software design in detail. Based on the characteristic of dual-core of DM6437 processor, the methods of communication protocol between ARM and DSP, as well as the double-buffer switching method are presented. Background modeling is made on the traffic video data from DM6437 via difference accumulation. Background subtraction is used to detect vehicle movement areas. With the adoption of algorithms such as Otsu, morphological filtering and region growing, the whole system is finally realized on the DM6446 hardware platform. Experimental results show that the system achieves good performance on moving vehicle detection.

Abstract: Using wireless sensors network (WSN) to collect coal mine hydraulic support pressure data , the method to solve the drawbacks of wired sensors . But the lack of WSN transmission performance in the coal mine : the transmission distance is short , large power consumption , short life cycle . Improve working methods for WSN : DIGI low-power sensor nodes ; using low-power single-chip Wake and Sleep MPS430149 at checking time ; core is a data fusion process , in the process of collecting data from all nodes , the use of local computing and storage nodes the ability to process the data , removing redundant data and minimize the amount of data transmitted within the network , improve data collection efficiency , achieve energy saving role in enhancing the network of life . The results show that the method is effective in improving the life and efficiency of WSN .

Abstract: In the absence of cooling measures, the chip temperature of high power LED rises rapidly. When the junction temperature exceeds the maximum allowable temperature, the high power LED will be damaged due to overheating. In the design of high power LED lamps, the main work is to the temperature design. Using a good linear relationship characteristic of the semiconductor devices between the LED junction voltage and junction temperature, this paper measured the high-power LED working voltage, working current, temperature of the incubator and heat sink. Adopting the method of electrical parameters, the paper measure the temperature resistor of high power LED. The high-power LED temperature resistance measuring instrument is designed and made. The temperature resistance from the LED wafer to the heat sink is measured fast and accurately, and the error is less than 5%.

Abstract: In this paper, the sensing principle of a B-dot probe to measure the armature position was analyzed. The output waveforms of the probe for two different orientations were given after mathematical deduction. An electromagnetic launching experiment was conducted, and the result shows that the output waveforms of the probes can present the velocity trend of the armature. After data fitting, the velocity of the armature moving in the bore was given, and combined with the waveform of the muzzle voltage, the muzzle velocity was obtained, which matched the measurement result by a high speed camera well.