Size Measurement Of High-temperature Components Based On Infrared Binocular Vision

Visible light vision measurement technique was commonly used as a kind of non-contact measurement method, but in the heat, smoke and other harsh environment, its image would be affected greatly, and the light deflection was serious. The basic working principle of infrared thermal vision technology was to receive the target’s radiation and complete the photoelectric conversion, which could overcome problems of the visible light vision.Centering on three-dimensional size measuring methods in harsh environments of high temperature, and in combination with the demand from a space research institute, a novel method based on infrared vision is proposed to measure size of high-temperature components, and the related theory and key techniques in this method are researched. The main research contents are as follows:1.The thesis summarizes related researches about wind tunnel test environment at home and abroad, and the application technological development of infrared thermal vision and binocular vision. The research direction and meaning are then determined.2. A model of infrared binocular vision measurement systems is established based on the ideal perspective model of infrared thermal imager. A measurement method combining a one-dimensional precision translation stage with single thermal imager is contrived. An overall scheme for the infrared vision measurement is contrived. And based on the actual situation, hardware devices are selected. Finally, the experiment platform is set up, which consists of the simulation system of tested components in high-temperature environment, the high-temperature infrared binocular vision measurement system and calibration system. Then some simple experiments are carried out.3.Non-uniformity correction method and image enhancement method of infrared image from high-temperature components is studied.4.Calibration method of internal and external parameters for infrared thermal imager in high-temperature measurements system is studied. A calibration method is designed to infrared high-temperature vision measurement combined with the characteristics of infrared thermal imaging. the measurement system is calibrated with a method which bases on one- dimensional radial alignment constraint (RAC). 5.The methods for edge extraction, binocular feature matching and size measurement are studied. An edge detection method based on fuzzy neural network and spatial gray moment is applied to effectively improve the adaptation and accuracy, and its precision is IR sub-pixel. The approaches to fitting initial matching based on polar constraint relationship and removing the mismatch of various methods are figured out. critical size measurement and shape reconstruction of high-temperature component are accomplished via data processing of 3D discrete coordinates.