Method for detecting remote sensing ground object target along railway based on dense network and attention mechanism

Abstract

The invention discloses a method for detecting a remote sensing ground object target along a railway based on a dense network and an attention mechanism, and relates to the field of deep learning and remote sensing image ground object target detection. The method comprises the following steps that a DenseNet module is used for replacing ResNet residual modules of a part of CSP units in a YOLOv4 network structure CSPDarknet53, so that feature reuse is achieved; a compression excitation structure is added in each CSPUnit in a backbone network of YOLOv4, so that the feature extraction capability is enhanced; a channel and a space attention mechanism are introduced before the network is output, so that the detection accuracy is improved; and a railway line remote sensing ground object target detection data set is manufactured, the improved YOLOv4 network structure is trained on the data set to obtain a trained railway line remote sensing ground object detection model, and ground object detection is carried out. The method can improve the detection speed and the detection precision, reduces the size of the model, is suitable for remote sensing ground object target detection along a railway, and meets the real-time requirement.

Description

technical field [0001] The invention belongs to the technical field of deep learning remote sensing image object detection technology, and specifically relates to a remote sensing object detection method along a railway based on a dense network and an attention mechanism. Background technique [0002] The geographical environment along the high-speed railway is complex, and there are many potential safety hazards. In particular, illegal buildings such as houses, pools, tunnels and factories along the line will seriously affect the operation safety of the high-speed railway. Therefore, it is necessary to conduct a timely investigation of the ground objects along the railway. The traditional manual inspection method is time-consuming, laborious and inefficient, and it is difficult to conduct a comprehensive inspection of the entire railway network. High-resolution remote sensing technology has technical advantages such as real-time and periodicity, and provides an effective t...

AI Technical Summary

Problems solved by technology

However, the above target detection algorithms are mainly used for the detection of conventional image targets. For remote sensing image targets with multiple bands and complex backgrounds, they cannot show their advantages in terms of accuracy and speed.

Images