矿用电缆火灾图像识别方法

Image recognition method of mine cable fires

针对目前矿井火灾的预测预报主要是依据温度、烟雾、CO、C02等参数进行判别,特别是对矿用电缆火灾监测基本上没有新的方法,并且上述方法在监测的准确性和有效性方面尚不能满足煤矿安全生产的需要。提出了采用图像法对矿用电缆火灾进行预测预报,介绍了矿用电缆火灾图像的处理方法,包括图像剪切、图像分割、灰度化处理、边缘轮廓提取、有效区域填充和面积计算,并用VC++进行应用软件设计。实践证明,图像处理和识别方法合理,效果较好,该方法的研究和应用克服了过去监测方法带来的弊端,对矿井火灾监测预报具有十分重要的理论意义和实用价值。

At present, mine fires are forecast with temperature, smog, CO, CO2, and other parameters. There is no new method to monitor mine cable fires. The above monitoring methods can＇t meet the requirements for safe production of coal mines in accuracy and validity of monitoring. The image method to forecast mine cable fires is put forward, which might solve past problems of monitoring methods. The authors have introduced the processing method of mine cable fires, which includes the image cut, the image segmentation, gradation processing, the edge contour extraction, the effective district filling and the area computation. And the authors use VC＋＋ for software design. Practice proves that the image processing and recognition method is reasonable and the effect is better. The research and application of this method overcomes the defects of the past monitoring methods and has great theoretical and practical value to forecast mine fires.