Improved FCOS Remote Sensing Image Detection Method Based on Distance Constraint

doi:10.3778/j.issn.1002-8331.2205-0478

Abstract

Abstract: Remote sensing image object detection method based on deep learning is usually difficult to eliminate background interference in complex scenes, which leads to low detection accuracy. In order to solve this problem, this paper designs a feature pyramid structure based on scale stratification, and proposes a distance-constraints centerness（DCCN）, thus forming an improved FCOS remote sensing image detection method based on distance constraint. The feature pyramid structure based on scale stratification includes high-level semantic information activation module and low-level effective feature perception module. The high-level semantic information module reconstructs the processing method of high-level feature map in the feature fusion stage, and improves the semantic perception ability of the top area of the feature pyramid. The low-level effective feature perception module enhances the information interaction ability between channels by introducing channel attention mechanism. DCCN can use the distance factor between the prediction box and the groundtruth box as the regression evaluation condition to improve the regression effect of the prediction box. In the experiment on NWPU VHR-10 dataset, the accuracy of this method reaches 92.6%, which is 4.9 percentage points higher than that of the original FCOS method, and effectively improves the accuracy of remote sensing image detection.

Key words: remote sensing image, object detection, centerness, attention mechanism

摘要： 基于深度学习的遥感图像目标检测方法通常难以排除复杂场景下的背景干扰，从而导致检测精度低。为解决该问题，设计了一种基于尺度分层的特征金字塔结构，并提出了一种基于距离约束的中心回归（distance-constraints centerness，DCCN），从而形成了基于距离约束的改进FCOS遥感图像检测方法。基于尺度分层的特征金字塔结构包括高层语义信息激活模块和低层有效特征感知模块，其中高层语义信息模块重构了特征融合阶段对高层特征图的处理方式，提升了特征金字塔顶部区域的语义感知能力，低层有效特征感知模块通过引入通道注意力机制，增强了通道间的信息交互能力。DCCN能够利用预测样本框与真实样本框之间的距离因素作为回归评估条件，提升了预测框的回归效果。在NWPU VHR-10数据集的实验中，该方法的精度达到92.6%，相比于原FCOS方法提升了4.9个百分点，有效改善了遥感图像检测的精度。

关键词: 遥感图像, 目标检测, 中心回归, 注意力机制

SU Shuzhi, XIE Yuqi. Improved FCOS Remote Sensing Image Detection Method Based on Distance Constraint[J]. Computer Engineering and Applications, 2023, 59(10): 227-235.

苏树智, 谢玉麒. 基于距离约束的改进FCOS遥感图像检测方法[J]. 计算机工程与应用, 2023, 59(10): 227-235.

References

[1] 杨军，杨婉钰.改进尺度不变特征变换的遥感影像配准[J].测绘科学，2021，46（5）：84-94.
YANG J，YANG W Y.Improved SIFT algorithm for remote sensing image registration[J].Science of Surveying and Mapping，2021，46（5）：84-94.
[2] 韩松来，王钰婕，王星，等.多尺度PCA-HOG遥感异源图像匹配方法[J].国防科技大学学报，2022，44（1）：146-155.
HAN S L，WANG Y J，WANG X，et al.Remote sensing multi-modal image matching algorithm based on multi-scale PCA-HOG[J].Journal of National University of Defense，2022，44（1）：146-155.
[3] 胡正平，董淑丽，赵淑欢.多尺度局部区域响应累积的非滑窗快速目标检测方法[J].信号处理，2016，32（1）：37-45.
HU Z P，DONG S L，ZHAO S H.Fast object detection algorithm with non-sliding window based on accumulation of multi-scale local response[J].Journal of Signal Processing，2016，32（1）：37-45.
[4] 魏驰宇，刘蓉，刘明，等.改进FCOS的复杂场景口罩佩戴检测算法[J/OL].计算机工程与应用（2022-04-17）[2022-05-18].http：//kns.cnki.net/kcms/detail/11.2127.TP.20220415.1235.
002.html.
WEI C Y，LIU R，LIU M，et al.Improved algorithm of FCOS for complex scene mask wear detection[J/OL].Computer Engineering and Applications（2022-04-17）[2022-05-18].http：//kns.cnki.net/kcms/detail/11.2127.TP.20220415.1235.
002.html.
[5] 刘竞升，伍星，王洪刚，等.改进FCOS的二阶段SAR舰船检测算法[J].计算机工程与应用，2021，57（24）：144-151.
LIU J S，WU X，WANG H G，et al.Improved FCOS two-stage SAR ship detection algorithm[J].Computer Engineering and Applications，2021，57（24）：144-151.
[6] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[C]//Advances in Neural Information Processing Systems 28，2015：91-99.
[7] REDMON J，FARHADI A.YOLOv3：an incremental improvement[J].arXiv：1804.02767，2018.
[8] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//14th European Conference on Computer Vision.Cham：Springer，2016：21-37.
[9] LIN T Y，GOYAL P，GIRSHICK R，et al.Focal loss for dense object detection[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision，2017：2980-2988.
[10] QING Y H，LIU W Y，FENG L Y，et al.Improved YOLO network for free-angle remote sensing target detection[J].Remote Sensing，2021，13（11）：2171.
[11] GAO F，HE Y，WANG J，et al.Anchor-free convolutional network with dense attention feature aggregation for ship detection in SAR images[J].Remote Sensing，2020，12（16）：2619.
[12] XU L，PANG C，GUO Y，et al.Combinational fusion and global attention of the single-shot method for synthetic aperture radar ship detection[J].Remote Sensing，2021，13（23）：4781.
[13] 陈欣，万敏杰，马超，等.采用多尺度特征融合SSD的遥感图像小目标检测[J].光学精密工程，2021，29（11）：2672-2682.
CHEN X，WAN M J，MA C，et al.Recognition of small targets in remote sensing image using multi-scale feature fusion-based shot multi-box detector[J].Optics and Precision Engineering，2021，29（11）：2672-2682.
[14] 刘畅，朱卫纲.多尺度与复杂背景条件下的SAR图像船舶检测[J].遥感信息，2021，36（3）：50-57.
LIU C，ZHU W G.Ship detection in SAR imagery under multi-scale and complex-background conditions[J].Remote Sensing Information，2021，36（3）：50-57.
[15] 张佳欣，王华力.改进YOLOv3的SAR图像舰船目标检测[J].信号处理，2021，37（9）：1623-1632.
ZHANG J X，WANG H L.Ship target detection in SAR image based on improved YOLOv3[J].Journal of Signal Processing，2021，37（9）：1623-1632.
[16] XU D Q，WU Y Q.MRFF-YOLO：a multi-receptive fields fusion network for remote sensing target detection[J].Remote Sensing，2020，12（19）：3118.
[17] 岳冰莹，陈亮，师皓，等.基于改进RetinaNet的SAR图像目标检测方法[J].信号处理，2022，38（1）：128-136.
YUE B Y，CHEN L，SHl H，et al.Ship detection in SAR images based on improved RetinaNet[J].Journal of Signal Processing，2022，38（1）：128-136.
[18] ZHU M，HU G，ZHOU H，et al.A ship detection method via redesigned FCOS in large-scale SAR images[J].Remote Sensing，2022，14（5）：1153.
[19] ZHENG Z，WANG P，LIU W，et al.Distance-IoU loss：faster and better learning for bounding box regression[C]//Proceedings of the 34th AAAI Conference on Artificial Intelligence，2020：12993-13000.
[20] TIAN Z，SHEN C，CHEN H，et al.FCOS：fully convolutional one-stage object detection[C]//Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision，2019：9627-9636.
[21] MOHAMMED M A，ABDULKAREEM K H，MOSTAFA S A，et al.Voice pathology detection and classification using convolutional neural network model[J].Applied Sciences，2020，10（11）：3723.
[22] LIN T Y，DOLLáR P，GIRSHICK R，et al.Feature pyramid networks for object detection[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition，2017：2117-2125.
[23] 高宇歌，杨海涛，王晋宇，等.联合知识与CNN的遥感影像目标检测研究综述[J].计算机工程与应用，2021，57（18）：65-74.
GAO Y G，YANG H T，WANG J Y，et al.Review of remote sensing image target detection research combining knowledge and CNN[J].Computer Engineering and Applications，2021，57（18）：65-74.
[24] JIANG B，LUO R，MAO J，et al.Acquisition of localization confidence for accurate object detection[C]//Proceedings of the 15th European Conference on Computer Vision，2018：784-799.
[25] HENDRYCKS D，GIMPEL K.Gaussian error linear units[J].arXiv：1606.08415，2016.
[26] GLOROT X，BORDES A，BENGIO Y.Deep sparse rectifier neural networks[C]//Proceedings of the 14th International Conference on Artificial Intelligence and Statistics，2011：315-323.
[27] HU J，SHEN L，SUN G.Squeeze-and-excitation networks[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition，2018：7132-7141.
[28] LIN T Y，MAIRE M，BELONGIE S，et al.Microsoft COCO：common objects in context[C]//13th European Conference on Computer Vision.Cham：Springer，2014：740-755.
[29] ZHANG S，CHI C，YAO Y，et al.Bridging the gap between anchor-based and anchor-free detection via adaptive training sample selection[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：9759-9768.
[30] DUAN K，BAI S，XIE L，et al.CenterNet：keypoint triplets for object detection[C]//Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision，2019：6569-6578.
[31] XIAO Z，LIU Q，TANG G，et al.Elliptic Fourier transformation-based histograms of oriented gradients for rotationally invariant object detection in remote-sensing images[J].International Journal of Remote Sensing，2015，36（2）：618-644.