基于数字图像和SDBP的预混火焰燃烧状态识别

利用数字图像处理技术和最速下降反传算法（SDBP）网络对预混火焰的燃烧状态识别技术进行了研究。针对本生灯的层流预混火焰，利用视频采集卡记录火焰图像，通过数字图像处理得到火焰的位置、形状信息。利用光电池记录火焰的辐射强度信息。将典型的火焰燃烧状态如稳定燃烧状态、回火状态以及脱火等状态的位置、强度信息作为改进后的SDBP人工神经网络的输入信息进行训练，训练好的SDBP网络接受火焰特征的输入信息，就可以识别火焰燃烧的状态。实验结果表明：该神经网络可以判断出当前火焰的燃烧状态。

Pre-mixed Flame Recognition Based on Digital Image and SDBP Neural Network

Pre-mixed flame recognition based on digital image and SDBP(steepest descent backpropagation) neural metwork is studied in this paper.Flame images is recorded by the video card,and the location and the shape information of the flame is acquired by the digital image processing.The radiant intensity is detected by the electric eye.The information for some typical combustion states such as the steady combustion state,the flameout state and the backfire state is used to train the improved SDBP neural network. The trained SDBP neural network can discern the combustion state by the flame character information.Experimental results show the trained SDBP neural network can discern the combustion state.