《電子技術(shù)應(yīng)用》
您所在的位置:首頁(yè) > 其他 > 設(shè)計(jì)應(yīng)用 > 改進(jìn)型DSSD算法在道路損傷檢測(cè)中的應(yīng)用研究
改進(jìn)型DSSD算法在道路損傷檢測(cè)中的應(yīng)用研究
2021年電子技術(shù)應(yīng)用第12期
蘇 可1,郭學(xué)俊2,楊 瑩3,陳澤華2
1.太原理工大學(xué) 電氣與動(dòng)力工程學(xué)院,山西 太原030024; 2.太原理工大學(xué) 大數(shù)據(jù)學(xué)院,山西 晉中030600;3.山西省交通科技研發(fā)有限公司,山西 太原030006
摘要: 在自動(dòng)檢測(cè)中,由于道路損傷數(shù)據(jù)集存在小目標(biāo)損傷難檢測(cè)與類(lèi)別不平衡問(wèn)題,導(dǎo)致道路損傷檢測(cè)的準(zhǔn)確率低、虛假率高。為此,在DSSD(Deconvolutional Single Shot Detector)網(wǎng)絡(luò)模型的基礎(chǔ)上,提出一種結(jié)合注意力機(jī)制和Focal loss的道路損傷檢測(cè)算法。首先,采用識(shí)別精度更高的ResNet-101作為DSSD模型的基礎(chǔ)網(wǎng)絡(luò);其次,在ResNet-101主干網(wǎng)絡(luò)中添加注意力機(jī)制,采用通道域注意力和空間域注意力結(jié)合的方式,實(shí)現(xiàn)特征在通道維度上的加權(quán)與空間維度上的聚焦,提升對(duì)小目標(biāo)道路損傷的檢測(cè)效果;最后,為了減少簡(jiǎn)單樣本的權(quán)重,增大難分類(lèi)樣本的權(quán)重,使用Focal loss來(lái)提高整體的檢測(cè)效果。在Global Road Damage Detection Challenge比賽所提供的數(shù)據(jù)集上進(jìn)行驗(yàn)證,實(shí)驗(yàn)結(jié)果表明,該模型的平均精度均值為83.95%,比基于SSD和YOLO網(wǎng)絡(luò)的道路損傷檢測(cè)方法的準(zhǔn)確率更高。
中圖分類(lèi)號(hào): TP391.41
文獻(xiàn)標(biāo)識(shí)碼: A
DOI:10.16157/j.issn.0258-7998.211684
中文引用格式: 蘇可,郭學(xué)俊,楊瑩,等. 改進(jìn)型DSSD算法在道路損傷檢測(cè)中的應(yīng)用研究[J].電子技術(shù)應(yīng)用,2021,47(12):64-68,99.
英文引用格式: Su Ke,Guo Xuejun,Yang Ying,et al. Research on application of improved DSSD algorithm in road damage detection[J]. Application of Electronic Technique,2021,47(12):64-68,99.
Research on application of improved DSSD algorithm in road damage detection
Su Ke1,Guo Xuejun2,Yang Ying3,Chen Zehua2
1.College of Electrical and Power Engineering,Taiyuan University of Technology,Taiyuan 030024,China; 2.College of Data Science,Taiyuan University of Technology,Jinzhong 030600,China; 3.Shanxi Transportation Technology Research and Development Co.,Ltd.,Taiyuan 030006,China
Abstract: In the automatic detection, the road damage data set has the problems of difficult detection of small target damage and imbalance of categories, resulting in low accuracy and high false rate of road damage detection. For this reason, based on the DSSD(deconvolutional single shot detector) network model, a road damage detection algorithm combining attention mechanism and Focal loss is proposed. First of all, ResNet-101 with higher recognition accuracy is used as the basic network of the DSSD model. Secondly, an attention mechanism is added to the ResNet-101 backbone network, and the channel domain attention and spatial domain attention are combined to achieve the weighting of features in the channel dimension and the focus on the spatial dimension, and improve the detection effect of small target road damage. Finally, in order to reduce the weight of simple samples and increase the weight of difficult-to-classify samples, Focal loss is used to improve the overall detection effect. It is verified on the data set provided by the Global Road Damage Detection Challenge competition. The experimental results show that the average accuracy of the model is 83.95%, which is more accurate than the road damage detection method based on SSD and YOLO network.
Key words : road damage detection;DSSD target detection algorithm;small target detection;attention mechanism;category imbalance problem

0 引言

    道路建設(shè)是衡量國(guó)家現(xiàn)代化水平的重要指標(biāo)之一,我國(guó)道路交通網(wǎng)龐大復(fù)雜,道路養(yǎng)護(hù)問(wèn)題凸顯。如何從道路圖像快速準(zhǔn)確地檢測(cè)出損傷區(qū)域及類(lèi)型成為學(xué)者研究的熱點(diǎn)。

    隨著深度學(xué)習(xí)的快速發(fā)展,使用卷積神經(jīng)網(wǎng)絡(luò)(Convolution Neural Network,CNN)[1]自主地從數(shù)據(jù)集中提取相應(yīng)特征信息成為主流方法,如快速的R-CNN[2]、SSD[3]、YOLO[4]等。這些網(wǎng)絡(luò)能夠定位和識(shí)別圖中具有邊界框的對(duì)象,為復(fù)雜背景下道路檢測(cè)提供了有效的框架。




本文詳細(xì)內(nèi)容請(qǐng)下載:http://ihrv.cn/resource/share/2000003873。




作者信息:

蘇  可1,郭學(xué)俊2,楊  瑩3,陳澤華2

(1.太原理工大學(xué) 電氣與動(dòng)力工程學(xué)院,山西 太原030024;

2.太原理工大學(xué) 大數(shù)據(jù)學(xué)院,山西 晉中030600;3.山西省交通科技研發(fā)有限公司,山西 太原030006)




wd.jpg

此內(nèi)容為AET網(wǎng)站原創(chuàng),未經(jīng)授權(quán)禁止轉(zhuǎn)載。