基于模式识别的火灾烟雾探测中抗水雾干扰研究

水雾与火灾烟雾在颜色、形态以及面积扩散等方面均具有相似性,容易对基于图像识别的火灾探测技术形成干扰。在同一地点及环境下使用监控系统录制火灾烟雾和水雾的真实视频,从运动方向特征和纹理特征着手进行分析,并利用支持向量机(support vector machine,SVM)对视频进行火灾烟雾和水雾的识别。实验结果表明,火灾烟雾与水雾在运动方向特征及纹理特征中熵和对比度上有明显差异,火灾烟雾识别时使用运动方向特征及纹理特征既能准确的识别出火灾烟雾,又能极大减小疑似烟雾目标带来的干扰。

Study on Water Mist Interference in Fire Smoke Detection Based on Pattern Recognition

Water mist and smoke in color, shape and area of diffusion were similar and easy to be detected by image recognition technology based on the formation of fire disturbance.In this paper, the real video monitoring system of fire smoke and water mist was recorded in the same location and environment,the direction of motion features and tex-ture features are analyzed, and the support vector machine of fire smoke and water mist on video recognition is used. The experimental results show that the fire smoke and water mist in the direction of motion features and texture fea-tures of entropy and contrast have obvious difference,which the fire smoke can accurately be identified and the interfer-ence brought by suspected of smoke target can greatly be reduced.