Detection Based On Flame Image Processing Nitrogen Oxide Emissions

China is one of the world's largest coal producing and consuming nations. And the coal-burning boilers take an important position on industrial production. Therefore, the industrial boiler's economical and full combustion have an important influence upon national economic development,improving people's living and improving environment. Optimization of combustion and boiler's burning conditions can not only reduce burning materials and avoid exploding accidents, but also can effectively control the emissions of burning pollutions, so it can improve the safety of industrial boiler combustion and reduce air pollution. At present, the burning management of industrial boilers is out of date, the coal quality is poor, the characteristics of coal often change, the parameters'select is difficult and the accurate and reliable means of nitrogen oxides detection are lacked. So thesis supplies the nitrogen oxide detection through flame image processing technology based on an improved PLS (PLS-Partial least-squares), the purpose is to dynamically detect the boiler's emissions of nitrogen oxides using image processing and give the parameters of them then give the simple, effective operation guidance information.Thesis studied several methods of flame image segmentation and the temperature extraction as well as nitrogen oxides detection methods. What's more, it also analyzes its characteristics. In allusion to the questions of image temperature extraction resulted from OTSU threshold as well as the slow speed in PLS detection of nitrogen oxides, the research gain the flame image in boiler through the CCD camera. Then use pseudo-color enhancement OTSU threshold segmentation (OTSU) and so on to segments the flame image. Using two-color method to acquire the reconstruction of temperature field and to extract the flame temperature, this method of acquiring the temperature is feasible. It can be applied to the diagnosis and analysis of the burning flame. Thesis focused on the mechanism of generation about nitrogen oxides and the influences of the running parameters on nitrogen oxides emission. This influence is the theory basis on detecting the emissions of nitrogen oxides. And the paper improve the calculating model on partial least squares, combined with the temperature extracted from the flame of the boiler combustion and the other parameters effecting the generation of nitrogen oxides, the paper predict emissions of nitrogen oxides and improve the accuracy and computing speed on forecast.In the end, the paper has summarized the research works and analyses the existing problems about the research. And it also conceives the automatic adjustment system on boiler in the industrial development.