Papers by Keyword: Detection Model

Abstract: Detection model of knotty illicit image is designed to detect special erotic image, such as nude chat image and advertising prostitution image, which are difficulty identified by conventional algorithm. Text information and skin-color information are detected in the image detection windows set according to the characteristic of images. Experimental results demonstrate that the model can obtain good results in detecting special erotic image.

Abstract: In view of using only the intensity information of retroreflection light and poor capacity of anti-interference in laser short-range detection, this paper presents a short-range detection method based on laser polarization properties of targets, designs the polarization detection model and constructs the polarization detection system. Through the analysis and the comparison of data which is obtained from the polarization characteristic and intensity characteristic experiments of five different targets, the conclusion that the detection method based on the laser polarization characteristics can effectively distinguish target is achieved. Meanwhile, the analysis of detection model in this paper can also provide certain theory significance for further research on polarization detection.

Abstract: To Resolve the garbage tag issue in Folksonomy, Lssvm algorithm for social spam detection model (least Squares support vector machine classifiers) was proposed. The method of inequality change the constraints in the traditional support vector machine into equality constraints, and take the empirical function of the squared error loss function as the Experience function in training set. so that the quadratic programming problem convert QP into solving linear equations, it was improving solution the speed of solution and accuracy of convergence.The experimental results show that we have got higher classification accuracyand less predict time than traditional svm detection methods based on least squares support vector machine algorithm garbage tag detection model.

Abstract: Tracking data containing outliers would seriously distort the accuracy of data processing for space tracking ships. Effective measures have to be adopted for the detection and correction of outliers. The time series of observations obtained by space tracking ships contain complex periodical elements caused by ship deformation and rolling. It brings difficulty to the detection of outliers. To solve the problem, a variety of detection methods were studied based on measurement elements representation model. The paper gives an analysis about the characteristic of measurements taken by the ships and proposes to detect the information about outliers by considering the correlation in time, space and data properties. A detection model based on spline function representation model is constructed and tested. Numerical results show that the proposed method can detect and correct outliers effectively, and provide an alternative to the current method of the measurements processing for the space tracking ships.

Abstract: Through research and analysis of the detection system plan demonstration system from the topology, software architecture and the basic systems functions, this paper discusses the basic components, the model library and run the process of scenario demonstration engine of plan demonstration in the detection system. It introduces the running control algorithms of scenario demonstration engine, the obscured corner priority judgment method, the command processor, scenario processor, and model design to achieve some of the key technology in scenario demonstration engine.

Abstract: The gauge detection between high signal and contact net is an important part in railway safety production. According to the detection nature, this paper designed a visual detection system to solve the problem of gauge detection between high signal and contact net. Using template matching method to break up the gray target image, and using discovering turning point method to rebuild target in space. Then according to the principle, it prepared detection system based on VC++2008 and OpenCV. The experiment results showed that: measurement error was within士15mm, when detection distance was arrange from 45.0m to 55.0m. Detection system which had high accuracy and good reliability, met system design requirements.

Abstract: It plays an important role to detect gauge between high signal and contact net in railway safe, to provide safe guarantee for railway workers. This paper proposed a visual detection model to solve the problem of gauge between high signal and contact net. Base on three-dimensional reconstruction high signal and contact line with space points, it transformed two objects coordinate, then analyzed the position relationship and accurately solved the shortest distance from high signal to contact line, last compared with the standard gauge to achieve the process of gauge detection.

Abstract: An aircraft cable fault location method based on detection model is proposed to solve the problem of being difficult to inspect the fault for the civil aviation maintenance. In response to the condition of the experimental installation, the reference signal is designed. The fault of the cable can be located according to the reflected waveform. An aircraft cable fault location system is designed and the experimental results show that the method is rational and effective.