The Safety , economy and the affection to the environment are the most important problems for the owners of thermal plants . In boilers , the combustion of the coal powder is directly connected to the operation . When unstable , it would reduce the efficiency , produce more pollution . And even in some conditions , the boilers may explode . So in the past time , many methods are used to insure the safe operation . As the development of digital image processing technology and computer technology , the combustion diagnosis based on the images is increasingly used in this field , which can directly and quickly provides the information about the combustion .The purpose , in this paper , is providing some information and stategies for the safe operation and combustion diagnosis through the analysis and processings of the flame images . We introduced some common digital image processings and intelligent theories , then computed some characteristics . In the last part of this paper , we trained a BP neural network with some trained samples . The network can achieves satisfied results with diagnosis sample inputs. |