Research on Explosion Point Image Recognition Based on Multimode Neural Network in Complex Ground Scene

Aiming at the problem that it is difficult to recognize the explosion point image in the complex ground scene, by analyzing the impact of the complex ground environment scene on the explosion point image recognition, a multi-mode fusion neural network architecture for image recognition suitable for complex background is constructed based on the algorithms such as YOLO and Mask R-CNN. On this basis, data enhancement technology and transfer learning technology are proposed to optimize the performance of the model. The results show that, compared with the traditional explosion point image recognition method, the multi-mode fusion neural network can effectively identify the explosion point in the complex ground scene, and the recognition accuracy, recall rate and F1 value have been improved to a certain extent, which can provide reliable technical support for the fire effect evaluation, situation analysis and fire planning in the complex ground scene.