Aiming at the bottleneck problem of insufficient adaptability of traditional defect detection methods in automated wood processing industry, research on intelligent detection technology based on deep learning is carried out, and a dataset covering multi-species wood characteristics and typical defect types is proposed. Applying object detection technology to defect detection, using dilation wise residual (DWR) module to optimize C2f module, and proposing task aligned dynamic detection head (TADDH) and feature focusing spread pyramid network (FSPN) to impove YOLOv8 algorithm (DFT-YOLO). The experimental results showed that a significant improvement in accuracy, reaching 96.8%, which was 7.9 higher than the original model. On the average accuracy of the key evaluation indicators mAP50 and mAP50-95, the impoved model reached 93.8% and 75.2%, respectively, increasing by 6.8% and 17.5%, respectively. While improving the detection accuracy, the number of parameters of the model had decreased by approximately 1/6 (16.2%). The impoved model can provide a lightweight detection method for wood defects.
为了评估每个模块的功效,通过消融试验对不同模块的缺陷检测结果进行了比较,比较方法包括了YOLOv8、YOLOv8+SI(single class improve,单一类别数据增强)、YOLOv8+SI+FSPN、YOLOv8+SI+FSPN+TADDH和YOLOv8+SI+FSPN+TADDH+DWR。表8为消融试验的结果。评估指标包括P、R、mAP50、mAP50-95和模型参数量大小。
CHENY, SUNC, RENZ,et al.Review of the current state of application of wood defect recognition technology[J].BioResources,2023,18(1),2288-2302.
[4]
URBONASA, RAUDONISV, MASKELIŪNASR,et al.Automated identification of wood veneer surface defects using faster region-based convolutional neural network with data augmentation and transfer learning[J].Applied Sciences,2019,9(22):4898.
[5]
DUX, LIJ, FENGH,et al.Image reconstruction of internal defects in wood based on segmented propagation rays of stress waves[J].Applied Sciences,2018,8(10):1778.
[6]
PUCCINIC T, GONÇALVESR, MONTEIROM E A.Avaliação estatística da variação da velocidade de propagação de ondas de ultra-som na madeira em presença de defeitos[J].Revista Brasileira de Engenharia Agrícola e Ambiental,2002,6:499-503.
[7]
LONGUETAUDF, MOTHEF, KERAUTRETB,et al.Automatic knot detection and measurements from X-ray CT images of wood:A review and validation of an improved algorithm on softwood samples[J].Computers and Electronics in Agriculture,2012,85:77-89.
[8]
LINW S, SHAOL J, WUJ Z.Study on wood board defect detection based on artificial neural network[J].The Open Automation and Control Systems Journal,2015,7(1):290-295.
MUH B, WANGS W, QID W,et al.Wood defects recognition based on gray-level co-occurrence matrix and fuzzy BP neural network[J].Forest Engineering,2017,33(4):40-43.
[11]
WANGD, LIUZ, CONGF.Wood surface quality detection and classification using gray level and texture features[C]//Advances in Neural Networks–ISNN 2015:12th International Symposium on Neural Networks,ISNN 2015,Jeju,South Korea,October 15-18,2015,Proceedings 12.Springer International Publishing,2015:248-257.
BAIX B, XUJ T, SONGE L,et al.Contrast of three image segmentation algorithms for wood surface defects[J].Journal of West China Forestry Science,2016,45(4):27-34.
[14]
LIUW, ANGUELOVD, ERHAND,et al.SSD:Single shot multibox detector[C]//Computer Vision–ECCV 2016:14th European Conference,Amsterdam,The Netherlands,October 11–14,2016,Proceedings,Part I 14.Springer International Publishing,2016:21-37.
[15]
REDMONJ, FARHADIA.YOLO9000:Better,faster,stronger[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Honolulu,HI,USA,November 9,2017.IEEE,2017:7263-7271.
[16]
BOCHKOVSKIYA, WANGC Y, LIAOH Y M.YOLOv4:Optimal speed and accuracy of object detection[J].arXiv preprint arXiv:2020.
[17]
REDMONJ, DIVVALAS, GIRSHICKR,et al.You only look once:Unified,real-time object detection[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition,Las Vegas,NV,USA,June 27-30,2016.IEEE,2016:779-788.
[18]
RENS, HEK, GIRSHICKR,et al.Faster R-CNN:Towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2016,39(6):1137-1149.
[19]
GIRSHICKR, DONAHUEJ, DARRELLT,et al.Region-based convolutional networks for accurate object detection and segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,38(1):142-158.
[20]
GIRSHICKR, DONAHUEJ, DARRELLT,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition,Columbus,OH,USA,September25,2014.IEEE,2014:580-587.
[21]
DINGF, ZHUANGZ, LIUY,et al.Detecting defects on solid wood panels based on an improved SSD algorithm[J].Sensors,2020,20(18):5315.
[22]
KURDTHONGMEEW, SUWANNARATK.Locating wood pith in a wood stem cross sectional image using YOLO object detection[C]//Proceedings of 2019 International Conference on Technologies and Applications of Artificial Intelligence (TAAI).Kaohsuing,Taiwan,January 16,2020.IEEE,2019:1-6.
JIANGX W, ZHAOX Q.Improved YOLOv7 algorithm for wood surface defect detection[J].Computer Engineering and Applications,2024,60(7):175-182.
[27]
WANGR, CHENY, LIANGF,et al.BPN-YOLO:A novel method for wood defect detection based on YOLOv7[J].Forests,2024,15(7):1096.
[28]
ANH, LIANGZ, QINM,et al.Wood defect detection based on the CWB-YOLOv8 algorithm[J].Journal of Wood Science,2024,70(1):26.
[29]
GIRSHICKR, DONAHUEJ, DARRELLT,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition,Columbus,OH,USA,June 23-28,2014.IEEE,2014:580-587.
[30]
HEK, GKIOXARIG, DOLLÁRP,et al.Mask R-CNN[C]//Proceedings of 2017 IEEE International Conference on Computer Vision.Venice,Italy,October 22-29,2017.IEEE,2017:2961-2969.
BOCHKOVSKIYA, WANGC Y, LIAOH Y M.YOLOv4:Optimal speed and accuracy of object detection[J].arXiv preprint arXiv:2020.
[33]
ZHANGY, LIUS, CAOJ,et al.A rapid,automated flaw segmentation method using morphological reconstruction to grade wood flooring[J].Journal of Forestry Research,2014,25(4):959-964.
[34]
CHANGZ, CAOJ, ZHANGY.A novel image segmentation approach for wood plate surface defect classification through convex optimization[J].Journal of Forestry Research,2018,29(6):1789-1795.
[35]
GUI Y H, ANDERSSONH, VICENR.Wood defect classification based on image analysis and support vector machines[J].Wood Science and Technology,2010,44:693-704.
[36]
KODYTEKP, BODZASA, BILIKP.A large-scale image dataset of wood surface defects for automated vision-based quality control processes[J].F1000Research,2021,10:581.
[37]
FENGC, ZHONGY, GAOY,et al.TOOD:Task-aligned one-stage object detection[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV).Montreal,QC,Canada,October 10-17,2021.IEEE,2021:3490-3499.
[38]
WEIH, LIUX, XUS,et al.DWRSeg:Rethinking efficient acquisition of multi-scale contextual information for real-time semantic segmentation[J].arXiv preprint arXiv:2022.