Research On Edge Detection Of Combustion Flame Image

By monitoring the flame of the boiler combustion, we can ensure the stability, security and economy of the boiler. Edge detection can provide parameters like the shape of the flame, position and size which are significant to the detection. The conventional edge detection operators, such as Roberts, Sobel, Canny, are linear filtering method with thedrawbacks of incomplete edge detection and poor anti-noise performance. Mathematical morphology is widely used in image detection. It is a nonlinear method founded on set theory with complete theoretical basis.Above all in this paper, the basic theory of mathematical morphology is discussed, its application in image processing is analyzed and introduced as well. Simultaneously gives the experimental results of operations to analyze the advantages and disadvantages. Based on the conventional morphology edge detection operators, this paper improves and amends the existing anti-noise operators. Combined with multi-scale concept and multi-structure elements, this article put forward a new anti-noise morphology edge detection algorithm of multi-scale concept and multi-structure elements.Finally, for the problem of the flame image detection, this paper experiments with the anti-noise morphology edge detection algorithm of multi-scale concept and multi-structure elements. The simulated result demonstrates that this operator has a good performance of noise reduction compared with traditional edge detectors. Meanwhile, the detected edge is smooth. So this method can be widely used in the later edge detection.