Junchao Zhu

An efficient image matching method using Speed Up Robust Features

The kernel of stereo vision system is stereo matching. In this paper proposed, an efficient image matching algorithm based on Speed Up Robust Features (SURF). Compared with traditional methods, this approach has enormous advantages of less computations and short time-consuming. The approach uses SURF algorithm to extract feature points. The feature descriptor of the approach is determined by Haar wavelet. The RANdom SAmple Consensus (RANSAC) algorithm is used to eliminate the false matches and wrong match point. The experimental results show that the approach is robust and has fast matching speed. That makes an important application field of 3D reconstruction.

Rub-impact fault diagnosis of rotating machinery based on Hilbert-Huang transform

The rub-impact fault is one of the important faults of the rotating machinery. The rub-impact fault signal may be nonlinear and non-stationary. The Hilbert-Huang transform (HHT) is based on time-frequency signal processing technology, which is suitable for nonlinear and non-stationary signals. In this paper, the HHT transform is used to analyze the time-frequency characteristic of the rub-impact fault signal of a rotor test rig. The velocity sensor is used to measure the vibration signal of the rub-impact fault rotor. The experimental result shows that the HHT transform can effectively detect the rub-impact fault of the rotor, which implies great potential of the HHT transform in the fault diagnosis of rotor system.

Study on multiple image processing hardware system based on DSP

This paper presents the hardware system scheme of multi-channel image processing based on DSP, with DSP as the core control of FPGA and CMOS, responsible for the overall scheduling system, using FPGA as coprocessor responsible for the collection of the original image information from CMOS. This paper combines the fisheye lens to obtain the complete original image for panoramic vision and stereo vision requirement. FPGA is used to design the SRAM cache controller based on Ping-pong mechanism. The data exchanging between DSP and FPGA are completed in EDMA model of DSP. Asynchronous FIFO is applied to solve the problem of cross clock domain. Experiments show that the system can be realized in the image data acquisition speed of 45 frames per second and the output of PAL style distortion fisheye image.

Instantaneous Characteristics of Nonlinear Torsion Pendulum and Its Application in Parameter Estimation of Nonlinear System

The nonlinear model of torsion pendulum is presented by considering the nonlinear damping force and nonlinear restoring force. The analytic solution of the nonlinear model is calculated to analyze the relationship between the characteristics of torsion pendulum and the nonlinear factors. The instantaneous characteristics of nonlinear torsion pendulum are analyzed by instantaneous undamped natural frequency and instantaneous damping coefficient. The instantaneous characteristics can be used for the parameter estimation of nonlinear torsion pendulum system. The nonlinear characteristics of the torsion pendulum are validated by the torsion pendulum based on the air-hovered turntable. The parameter estimation method based on the instantaneous characteristics is validated by the moment of inertia measurement system based on the torsion pendulum.

Intelligent monitoring system of light intensity and CO 2 concentration in strawberries greenhouse

Aiming at the lack of stability and precision of the traditional greenhouse monitoring system, we constructed the intelligent monitoring system to monitor the light intensity and CO2 concentration of strawberries greenhouses that is based on the STM32 single-chip, by which the light intensity and CO2 concentration can be acquired and displayed at the same time. The serial communication is used to send the detection results to the MCGS touch screen, then the real-time monitoring and the alarm of the light intensity and CO2 concentration in the greenhouse can be realized. The long-term test results show that the detection accuracy of the light intensity in this system is ± 940Lux, the detection accuracy of the CO2 concentration is ± 40ppm. Compared with the traditional automatic monitoring system, the system runs stably and has low cost, whose precision can meet the requirements of the strawberries greenhouse. Above all, the system presented in this paper is suitable for applications to agricultural fields.

Feature matching method for uncorrected fisheye lens image

Traditional matching algorithms cannot be directly applied to the fisheye image matching for large distortion existing in fisheye image. Therefore, a matching algorithm based on uncorrected fisheye images is proposed. This algorithm adopts a local feature description method which combines MSER detector with CSLBP descriptor to obtain the image feature. First, the two uncorrected fisheye images captured by binocular vision system are described by the principle of epipolar constraint. Then the region detection is done with MSER and the ellipse fitting is used to the obtained regions. The MSER regions are described by CSLBP subsequently. Finally, in order to exclude the mismatching points of initial match, random sample consensus (RANSAC) algorithm has been adopted to achieve exact match. Experiments show that the method has a good effect on the uncorrected fisheye image matching.

Error analysis of fisheye correction curve

Various types of image can be captured with fisheye lens, their wide field of view is particularly suited to a stereo vision. However, fisheye lens introduces distortion, this change makes the image difficult to identify. If a correction curve can be fitted to the distortion shape, the degree of distortion can be described, support vector machine is a popular machine learning method for classification and regression, support vector classification can find optimal interval of the different classified data, the classified curve is as the correction curve, the training data is obtained by corner detection, the center of the data is gained by Hough transform. Although ignoring the error in the process of the algorithm, but the error is still existed during the picture taken from a fisheye lens in different situation, for example angle or distance, then using the support vector regression to gain the error between the original data and predicted data, original data stem from training target which is also obtained by a fisheye lens and predicted data stem from another fisheye image.

Design and hardware implement of automatic white balance based on chromaticity difference

This paper tries to solve the problem that different degrees of color cast will be presented when collects images from various directions in the same circumstance by using the system of Multi-vision. An automatic white balance algorithm based on the chromaticity difference is proposed, in which adopts the color difference formula CIE 1976Lab, comparing the Chromatic difference with threshold value of adjacent pixels and removing the interference pixels, the color temperature will be more accurate. And using integer multiplies and shift operation instead of floating-point arithmetic, the transformation from RGB color system to XYZ color space is optimized. This paper will start with description of algorithm principles. Following this, Chromatic difference white balance module will be built combining with DSP builder, and using Signal Compiler, the algorithm will be implemented by transplanting codes into FPGA. At last, Experimental results show that the proposed method can get a comparatively good color white balance effect by applying this algorithm into the latest Multi-vision system of panoramic stereo sphere vision.

Research on image matching technology for the spherical stereo vision

The technique of stereo matching is an important research goal of spherical stereo vision system, which is widely used in visual navigation, motion analysis and panoramic surveillance areas. Fisheye lens can get view images of large field, this paper uses two fisheye lens whose optical axis are perpendicular to construct spherical stereo vision system and does research in stereo matching of spherical stereo vision system. As the image that fish eye lens collected has lager distortion, this paper does stereo matching with the images the system collected without distortion correction using matching algorithm based on maximally stable extremal regions (MSER) and affine scale invariant feature transform (ASIFT) to reduce the correction error generated in the process. The experiment results show that the algorithm can be applied in stereo matching of fish eye image without distortion, and the results meets the need of 3D reconstruction.

Research on calibration method of the panoramic stereo sphere vision system

Panoramic stereo spherical vision system can simultaneously realize panoramic stitching and three-dimensional ranging. Its multi-channel image acquisition system is located in 3D space through the mechanical encapsulation. In this paper, an effective calibration method is explored to solve the problem of slight errors generated by the encapsulation structure installation and manufacturing. More importantly, the errors are difficult to measure, which leads the optical axes of lenses are not parallel. This paper uses a fine-tuning method of focused laser spot detection to solve the problem by analyzing the encapsulation structure and imaging model of the system. And its image processing algorithms and calibration process are described. The final experimental results confirm it is indeed feasible to this method.