摘要
探索普通日光条件下烟雾的彩色、时空特征的选择方法,通过Relief、PCA等特征选取方法选择适应烟雾探测并在不同彩色空间和变换空间下的各通道及分量和运动的最佳分类特征顺序,用协方差矩阵验证各颜色特征分类的可区分性贡献率,通过暗通道优先获得烟雾区域特征,将其作为二次候选探测区域的验证方法,实验结果表明,选择出的烟雾彩色融合特征在探测系统中表现出较高的识别精度和运行效率。
The selecting method for smoke’s colors and spatial-temporal features was probed,and a method of producing optimal classification feature of the sequence based on degree-of-contribution for optimal classification to these candidate color and moving subsets by Relief analyzing and PCA transformation.Contribution ratio of discrimination of all selected colors and moving fea-tures were validated using covariance matrix.The verifying methods for second candidate-region by drawing smog area feature according to dark channel prior were probed.Experimental results show that the smoke detection system based on color and moving feature has higher recognition accuracy and efficiency.
出处
《计算机工程与设计》
北大核心
2016年第7期1867-1872,共6页
Computer Engineering and Design
基金
国家自然科学基金项目(61262031)
