改进CenterNet的无人机小目标捕获检测方法

doi:10.3778/j.issn.1002-8331.2111-0033

摘要/Abstract

摘要： 小目标因浅层特征语义不足而深层特征信息缺失导致极难检测，而无人机视角场景复杂，检测难度进一步增大。普遍提升小目标检测精度的方法是进行不同层级的特征融合，但这会导致特征高冗余问题，并非所有特征层都值得被激活传递到后方的数据预测中去。针对上述问题对CenterNet进行改造，首次将其与自适应特征激活相结合，提出自适应基础模块（MSA），抑制冗余特征的表达；在主干输出处引入升维全局上下文注意力模块（GC-Block），强化关键点语义信息；用深度可分离卷积与Mish激活搭建高质量解码块（DW），在不增加模型复杂度的情况下提升解码精度。在公开的无人机捕获小目标数据集上进行对比实验，改进算法的AP较原始算法提升了2.2个百分点，召回率提升了2.4个百分点，验证了改进算法对小目标检测任务的有效性。

关键词: 无人机捕获目标检测, 小目标检测, anchor-free, 自适应激活, 注意力机制

Abstract: Small targets are extremely difficult to detect due to insufficient low-level semantics and lack of high-level feature information. The scene of the UAV’s perspective is complex, which further increases the difficulty of detection. The general method to improve the detection accuracy of small targets is to perform feature fusion at different levels, but it will cause the problem of high feature redundancy. Not all feature layers are worthy of being activated and passed to the rear data prediction. To solve the above problem, CenterNet is modified, combined with adaptive feature activation for the first time, and modified self-adaptive basic block is proposed to suppress the expression of redundant features；secondly, dimension raising global context block is introduced at the output of backbone to strengthen the key point semantic information；finally, deep separable convolution and Mish activation are used to build a high-quality decoding block abbreviated as DW, to improve the decoding accuracy without increasing the complexity of the model. Comparing experiments on the public UAV capture small target datasets, the AP of the improved algorithm is 2.2 percentage points higher than the original algorithm, and the recall is increased by 2.4 percentage points, which verifies the effectiveness of the improved algorithm for small target detection tasks.

Key words: UAV capture target detection, small object detection, anchor-free, adaptive activation, attention mechanism

刘鑫, 黄进, 杨涛, 王晴. 改进CenterNet的无人机小目标捕获检测方法[J]. 计算机工程与应用, 2022, 58(14): 96-104.

LIU Xin, HUANG Jin, YANG Tao, WANG Qing. Improved Small Object Detection for UAV Acquisition Based on CenterNet[J]. Computer Engineering and Applications, 2022, 58(14): 96-104.

参考文献

[1] ZHAO Z Q，ZHENG P，XU S，et al.Object detection with deep learning：a review[J].IEEE Transactions on Neural Networks and Learning Systems，2019，30（11）：3212-3232.
[2] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//IEEE Conference on Computer Vision and Pattern Recognition，2014.
[3] GIRSHICK R.Fast R-CNN[C]//IEEE International Conference on Computer Vision，2015.
[4] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis & Machine Intelligence，2017，39（6）：1137-1149.
[5] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[J].arXiv：1512.02325，2015.
[6] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[7] REDMON J，FARHADI A.YOLO9000：better，faster，stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：7263-7271.
[8] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[9] LIN T Y，GOYAL P，GIRSHICK R，et al.Focal loss for dense object detection[J].IEEE Transactions on Pattern Analysis & Machine Intelligence，2020，42（2）：318-327.
[10] LAW H，DENG J.Cornernet：detecting objects as paired keypoints[C]//Proceedings of the European Conference on Computer Vision（ECCV），2018：734-750.
[11] TIAN Z，SHEN C，CHEN H，et al.Fcos：fully convolutional one-stage object detection[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision，2019：9627-9636.
[12] ZHOU X，ZHUO J，KRAHENBUHL P.Bottom-up object detection by grouping extreme and center points[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：850-859.
[13] DONG Z，LI G，LIAO Y，et al.Centripetalnet：pursuing high-quality keypoint pairs for object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：10519-10528.
[14] ZHOU X，WANG D，KR?HENBüHL P.Objects as points[J].arXiv：1904.07850，2019.
[15] ZHOU X，KOLTUN V，KR?HENBüHL P.Probabilistic two-stage detection[J].arXiv：2103.07461，2021.
[16] 梁鸿，王庆玮，张千，等.小目标检测技术研究综述[J].计算机工程与应用，2021，57（1）：17-28.
LIANG H，WANG Q W，ZHANG Q，et al.Small object detection technology：a review[J].Computer Engineering and Applications，2021，57（1）：17-28.
[17] LIN T Y，DOLLAR P，GIRSHICK R，et al.Feature pyramid networks for object detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2017.
[18] LIU S，QI L，QIN H，et al.Path aggregation network for instance segmentation[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition（CVPR），2018.
[19] LIU S，HUANG D，WANG Y.Learning spatial fusion for single-shot object detection[J].arXiv：1911.09516，2019.
[20] GHIASI G，LIN T Y，LE Q V.NAS-FPN：learning scalable feature pyramid architecture for object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：7036-7045.
[21] TAN M，PANG R，LE Q V.Efficientdet：scalable and efficient object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：10781-10790.
[22] QIAO S，CHEN L C，YUILLE A.Detectors：detecting objects with recursive feature pyramid and switchable atrous convolution[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2021：10213-10224.
[23] CAI Z，FAN Q，FERIS R S，et al.A unified multi-scale deep convolutional neural network for fast object detection[C]//European Conference on Computer Vision.Cham：Springer，2016：354-370.
[24] HAN K，WANG Y，TIAN Q，et al.Ghostnet：more features from cheap operations[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：1580-1589.
[25] YU F，WANG D，SHELHAMER E，et al.Deep layer aggregation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：2403-2412.
[26] NEWELL A，YANG K，DENG J.Stacked Hourglass networks for human pose estimation[C]//European Conference on Computer Vision.Cham：Springer，2016：483-499.
[27] REZATOFIGHI H，TSOI N，GWAK J Y，et al.Generalized intersection over union：a metric and a loss for bounding box regression[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：658-666.
[28] MA N，ZHANG X，LIU M，et al.Activate or not：learning customized activation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2021：8032-8042.
[29] CHOLLET F.Xception：deep learning with depthwise separable convolutions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：1251-1258.
[30] CAO Y，XU J，LIN S，et al.Gcnet：non-local networks meet squeeze-excitation networks and beyond[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops，2019.
[31] WANG X，GIRSHICK R，GUPTA A，et al.Non-local neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：7794-7803.
[32] HU J，SHEN L，SUN G.Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：7132-7141.
[33] ZHU P，WEN L，DU D，et al.Vision meets drones：past，present and future[J].arXiv：2001.06303，2020.
[34] WU Y X，KIRILLOV A，MASSA F，et al.Detectron2[EB/OL].[2021-09-20].https：//github.com/facebookresearch/detectron2.
[35] DU D，ZHU P，WEN L，et al.VisDrone-DET2019：the vision meets drone object detection in image challenge results[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops，2019.