Advanced Materials Research Vols. 712-715

Abstract: Target precise positioning by vision system is one of key techniques in robot vision system. In target positioning and selection with robot vision technique, the camera lens distortion must be calibrated. In this paper, a calibration method based on segment slope is used to calibrate the camera and the radial lens distortion coefficient is obtained. The distortion coefficient is used in calculating target position coordinates, and the robot end-exceutor is led to position the target with the use of the coordinates. The experimental results show the effectiveness of the research work. Keywords:robot vision;camera calibration;radial distortion;target positioning

Abstract: Detecting of the auto parts solder joints quality is always a technological difficulty in the production line, the key point is how to exact the parts feature which is tested. New quality detection method for the auto parts soldered joints is presented based on biology vision feature. First the Gabor filter banks which have the emulation of biological vision is used to filter the detected image with eight different directions. Then the biggest powers of all pixel points are chosen to be the solder joints feature and classified by support vector machine (SVM). The experimental results show that the algorithm has higher accuracy as an effective quality detection method. It can be a useful reference to the auto parts solder joints quality testing of engineering application.

Abstract: The paper presented a conventional sequential Structure from Motion method, introduced the algebraic concepts and the core techniques used effectively. It is especially essential to comprehend the static scene based sequential of Structure from Motion is extended to a simultaneous segmentation and reconstruction scheme for dynamic scenes. The first outcome is the multiple view geometry, the second is the feature tracking and geometry initializing, and the last is self-calibration.

Abstract: SIFT (scale invariant key points) which can well handle images with varying orientation and zoom, is widely used in image registration, but the algorithm is complexity and the processing time is too long. Therefore we used the PCA-SIFT (Principle Components Analysis-scale invariant key points) in image registration. Compared the SIFT descriptor, the PCA-SIFT reduced the dimensions of SIFT feature, enhanced the matching accuracy and reduce the elapsed time. Then the mutual information method used in this paper to estimate the best points. Experimental results show that PCA-SIFT algorithm is simplified, robust and liable.

Abstract: In this paper, An effective shot cut detection algorithm directly in DCT domain was proposed. This algorithm(including abrupt change and ) test the n -1 frame and n frame through the extracting the feature vector of each frame. When extract the features of frame, two-dimensional mathematics parameter m1-s from DCT coefficients without its inverse transform and 8 directional chain code was computed, Divide m1-s space into 30 unequal partitions (subspaces) and compute the numbers pr within the 30 subspaces (entries) as the feature vector to judge the different of the two frames. locating shot cuts is operated by comparison tests.

Abstract: Since how to evaluate the ship wake images quality quantitatively and effectively, and achieve the purpose of the extraction of ship wakes optical signal characteristics rapidly, is an important base task for the ship optics-wakes elimination technology now. The method about quality evaluation of ship wake images is studied in this paper based on the contrast model. According to the characteristics of the obvious contrast between wake area and seawater background area in ship wake images, this paper uses the contrast parameter as the index of wake image quality evaluation, and also as the characterization of the strength of ship wakes optical signal characteristic. An experiment about the extraction of contrast feature and its results show that the contrast parameters can not only reflect the subjective visual perception of the human eye on the wake images actual quality, but also better reflect the application essence of the wake image processing. And the contrast parameters can describe the intensitys changing rules of ship wakes optical signal characteristics reliably and quantitatively.

Abstract: Segmentation is an important part of fingerprint image preprocessing. Effective segmentation not only reduces the time of subsequent processing but also improves the reliability of feature extraction considerably. After introducing segmentation, we suggest improved mean and variance-orientation coherence gradual segmentation algorithm of fingerprint image. Morphology has been applied as preprocessing to reduce the number of classification errors. The algorithm is tested on FVC2002 database, only 0.79% of the blocks are misclassified, while the preprocessing further reduces this ratio. Experimental results verify the feasibility of this algorithm.

Abstract: A Local Piecewise-Linearity Prediction method is presented, Based on the advantages and limitations of local prediction of chaotic time series. Taking time series of rainfall as example for prediction the rainfall of one city in Liaoning province, which includes the application of the largest Lyapunov exponent, Local-region method and Local Piecewise-Linearity method. The method proposed is proved practical in comparison with the observed data.

Abstract: Inertial Navigation System (INS) and Global Positioning System (GPS) are commonly integrated to overcomes each systems inadequacies and provide an accurate navigation solution. However, the positioning accuracy deteriorates with time in the absence of GPS signals. This paper proposed an artificial neural network aided INS/GPS integrated navigation method to bridge the GPS outages. In this method, the neural network was trained to model the error of INS when GPS was available. INS velocity and attitude error were got by implementing the neural network when GPS outages occurred. Simulations in INS/GPS integrated navigation showed improvement in positioning accuracy for GPS outages.

Abstract: To overcome the shortage that the particle swarm optimization is prone to trap into local extremum searching for the lost in population diversity, a strategy in which the velocity is not dependent on the size of distance between the individual and the optimal particle but only dependent on its direction is proposed. The average similarity of particles in the population is seem as the measure of population diversity and it is used to balance the global and local searching of the algorithm. Based on establishing the relationship between inertia weight and the measure of population diversity which has been inserted into the algorithm, A resilient particle swarm optimization algorithm with dynamically changing inertia weight (ARPSO) was proposed. ARPSO was applied in simulation experiment. The results show that the algorithm has the ability to avoid being trapped in local extremum and advance the probability of finding global optimum.