Abstract:Object detection plays a vital role in remote sensing applications. Although object detection has achieved proud results in natural images, these methods are difficult to be directly applied to remote sensing images. Remote sensing images often have complex backgrounds and small objects, which results in a highly unbalanced distribution of foreground and complex background information. In order to solve the above problems, this paper proposes a multi-head channel and spatial trans-attention (MCSTA) module, which performs remote pixel interaction from the channel and spatial dimensions respectively to complete the attention feature capture function. It is a plug-and-play module that can be easily embedded in any other natural image object detection convolutional neural network, making it quickly applicable to remote sensing images. First, in order to reduce computational complexity and improve feature richness, we use a special linear convolution to obtain three projection features instead of the simple matrix multiplication transformation in Transformer. Second, we obtain trans-attention maps in different dimensions in a manner similar to the self-attention mechanism to capture the interrelationships of features in channels and spaces. In this process, we use a multi-head mechanism to perform parallel operations to improve speed. Furthermore, in order to avoid large-scale matrix operations, we specially designed an attention blocking mode to reduce computer memory usage and increase operation speed. Finally, we embedded the trans-attention module into YOLOv8, added a new detection head and optimized the feature fusion method, thus designing a lightweight small object detection model named TA-YOLO for remote sensing images. It has fewer parameters than the benchmark model YOLOv8, and its mAP on the PASCAL VOC and VisDrone data sets increased by 1.3% and 6.2% respectively. The experimental results prove the powerful function of the trans-attention module and the excellent performance of TA-YOLO.

Remote sensing object detection based on a combination of a CNN and the Swin transformer

Swin-Transformer-Based YOLOv5 for Small-Object Detection in Remote Sensing Images

A Lightweight SE-YOLOv3 Network for Multi-Scale Object Detection in Remote Sensing Imagery.

A Convolution with Transformer Attention Module Integrating Local and Global Features for Object Detection in Remote Sensing Based on YOLOv8n

An Improved Swin Transformer-Based Model for Remote Sensing Object Detection and Instance Segmentation

Swin-Transformer-Enabled YOLOv5 with Attention Mechanism for Small Object Detection on Satellite Images

C3TB-YOLOv5: integrated YOLOv5 with transformer for object detection in high-resolution remote sensing images

Object Detection Based on Swin Deformable Transformer-BiPAFPN-YOLOX

A Refined and Efficient CNN Algorithm for Remote Sensing Object Detection

An Efficient Object Detection Algorithm Based on Improved YOLOv5 for High-Spatial-Resolution Remote Sensing Images

An object detection algorithm combining self-attention and YOLOv4 in traffic scene

Adaptively Attentional Feature Fusion Oriented to Multiscale Object Detection in Remote Sensing Images

CNN-transformer mixed model for object detection

Remote Sensing Object Detection Based on Strong Feature Extraction and Prescreening Network

Oriented Object Detection Based on CNN and Transformer for Remote Sensing Images

Multimodal Transformer Using Cross-Channel attention for Object Detection in Remote Sensing Images

Remote Sensing Image Object Detection Based on Improved YOLOv5

TA-YOLO: a Lightweight Small Object Detection Model Based on Multi-Dimensional Trans-Attention Module for Remote Sensing Images

ST-YOLOX: a Lightweight and Accurate Object Detection Network Based on Swin Transformer

Application of local fully Convolutional Neural Network combined with YOLO v5 algorithm in small target detection of remote sensing image