基于多分类器融合的林火检测方法研究

为了解决现有视频火焰检测算法在环境发生变化时识别精准度低、检测结果不佳等问题，提出了一种基于DS证据理论的多分类器融合林火检测方法。该方法通过提取疑似区域，对比选取了颜色、圆形度、面积三种火焰特征，然后分别输入支持向量机(SVM)、最临近(KNN)和决策树(DT)中进行分类识别，最后利用DS证据理论进行决策级融合。通过与其他方法对比实验表明，该方法受环境变化的影响较小，当识别场景发生改变时，识别精准度变化不超过3%,仍保持较高的识别精准度，具有良好的应用前景。

Research on Forest Fire Detection Method Based on Multi-classifier Fusion

In order to solve the problems of low recognition accuracy and poor detection results of the existing video flame detection algorithms when the environment changes, a multi-classifier fusion forest fire detection method based on DS evidence theory is proposed. This method extracts the suspected area, compares and selects three flame characteristics, color, roundness and area, and then inputs them into Support Vector Machine (SVM), K Nearest Neighbor (KNN) and Decision Tree (DT) for classification and recognition. Finally, the DS evidence theory is used for decision-making fusion. Compared with other methods, the experiment shows that this method is less affected by environmental changes, and when the recognition scene changes, the recognition accuracy changes by less than 3%, and it still maintains a high recognition accuracy, so it has a good application prospect.