Inspection and Firefighting Robot based on Improved SVM Algorithm in Ship Automation Engine Room

With the increase of ship automation, the number of electrical equipment placed in the engine room also increases, which leads to a rapid spread of fire in the engine room. Therefore, in order to prevent accidents and reduce personal and property losses caused by fires, a flame recognition algorithm based on firefly algorithm and support vector machine is proposed for inspection and firefighting robots. The experimental results show that the improved firefly algorithm - Support Vector Machine has the highest recognition accuracy and recall rates of 98.7% and 97.2%, respectively, and the lowest are 98.1% and 96.3%, respectively. The average accuracy and recall rates are 98.4% and 96.7%, respectively; The highest accuracy and F1 mean are 97.9% and 0.979, respectively. The average accuracy and F1 mean are 97.6% and 0.975, respectively, which are higher than other algorithms. The above results indicate that the flame recognition algorithm based on firefly algorithm and support vector machine proposed in the study can accurately identify fires, providing strong support for the work of inspection and firefighting robots.