Research On The Online Detecting System For Vision Coordinate Of Thermal Abnormity Surface

This paper is supported by the key project of Anhui Province. . In this project, the system of online detecting vision coordinate of abnormal surface effectively and credibly under special situation such as high temperature is studied.As an application, this paper aims at the urgent problem of detecting thermal train-wheel. The parameters of train-wheel are obtained from one thermal image otherwise from more sensors.According to the need of system, a block diagram is designed, a mathematics model is setup and the construct parameters are optimized. Based on the images of high temperature component, an effect of the laser spectrum to restrain thermal background and extrude the aim is analyzed. Considering of the spectrum efficiency for human eyes and the sensitive of a CCD camera respond to the spectrum, a logical laser is selected. An optics route of laser collimation, expending and transforming the beam is optimized by use of optics software. The changed laser target has centralized energy and uniformity width.An optical filter is selected to filter the interferential light from thermal background and extrude the laser target, and also, reduce the interference of the radiant heat to image information and the influence from surrounding temperature to the cameras.According to the image characteristics of the high-temperature components, the spectrum components of the thermal images are analyzed in detail, a effective method and some parameters are selected. As the first innovative option, the weights are important to the gray value transform in the image process, and a weighting method of the gray value transform is putout by the author. In addition, another two methods of image process for high-temperature components are designed: the method of mapping transform with twice average and the method of the gray value transform with the spectrum components.The calibration of a camera system is a key step to the precise conversion between an image of machine vision and the real dimension. The effect and efficiency of some re-sampling calculations for image edge are discussed. A calibration methodof plane to plane is designed. The experiments show that the method is precise and simply.As the second innovative option in this paper, the high temperature field of a measured train wheel is analyzed by use of limited element method at first time. The data and curves for thermal deformation are obtained for some dimensions and shapes. The experiments show that the theory calculation is available and credible.As the third innovative option in this paper, the air temperature field around the measured train wheel is simulated by use of limited element method at first time. According to the distribution rule of air temperature and air refractive index, the thermal deformation of a light beam in some position is calculated. The influence of the high temperature air field to measurements is avoided by arranging the logical positions of the measuring sensors.Besides, the method of twice average threshold is putout for the mapping transform in the image process of high- temperature components. The measurable effect of image process for edge detecting is assessed at first time. The influence from surrounding temperature to the CCD cameras is analyzed and discussed in experiments at first time.All these works and results enrich the connotation of machine vision. A new thought is provided for detecting vision coordinate of abnormity surface under thermal condition. The research works in this paper are contributory to the industry measurements in theory and practical application.