基于改进FOA-SVM的火灾图像识别模型研究

FIRE IMAGE RECOGNITION MODEL BASED ON IMPROVED FOA-SVM

火灾是常见的破坏性极大的自然灾害。为了更好地预防火灾,减少财产损失和人员伤亡,针对人为选择SVM参数具有盲目性,对其分类能力影响较大,提出基于改进FOA-SVM的火灾图像识别模型。通过引入逻辑函数对果蝇算法的搜索步长进行改进,利用改进果蝇算法优化支持向量机搜索得到最佳模型参数。将火灾图像提取特征量作为该识别模型的输入样本训练和识别火灾图像,结合实例并将该模型的识别结果与SVM模型及其他算法的识别结果进行对比。实验结果表明,该模型提高了火灾图像识别的准确率,在火灾检测方面具有一定的实际应用价值。

Fires are common and devastating natural disasters.In order to better prevent fires,reduce property losses and casualties,aiming at the blindness of artificial selection of SVM parameters,which has a greater impact on its classification capabilities,we proposed a fire image recognition model based on improved FOA-SVM.The logic function was introduced to improve the search step of fruit fly algorithm,and we used the improved fruit fly algorithm to optimize the SVM to obtain the optimal model parameters.The fire image extraction feature quantity was used as the input sample of the recognition model to train and identify the fire image,then we combined the example and compared the recognition results of the model with the identification results of the SVM model and other algorithms.Experimental results show that the model improves the accuracy of fire image recognition and has certain practical application value in fire detection.