Surface Detection of Solid Wood Defects Based on SSD Improved with ResNet
Due to the lack of forest resources in China and the low detection efficiency of wood surface defects, the output of solid wood panels is not high. Therefore, this paper proposes a method for detecting surface defects of solid wood panels based on a Single Shot MultiBox Detector algorithm (SSD) to d...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-10-01
|
| Series: | Forests |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4907/12/10/1419 |
| _version_ | 1797514513706123264 |
|---|---|
| author | Yutu Yang Honghong Wang Dong Jiang Zhongkang Hu |
| author_facet | Yutu Yang Honghong Wang Dong Jiang Zhongkang Hu |
| author_sort | Yutu Yang |
| collection | DOAJ |
| description | Due to the lack of forest resources in China and the low detection efficiency of wood surface defects, the output of solid wood panels is not high. Therefore, this paper proposes a method for detecting surface defects of solid wood panels based on a Single Shot MultiBox Detector algorithm (SSD) to detect typical wood surface defects. The wood panel images are acquired by an independently designed image acquisition system. The SSD model included the first five layers of the VGG16 network, the SSD feature mapping layer, the feature detection layer, and the Non-Maximum Suppression (NMS) module. We used TensorFlow to train the network and further improved it on the basis of the SSD network structure. As the basic network part of the improved SSD model, the deep residual network (ResNet) replaced the VGG network part of the original SSD network to optimize the input features of the regression and classification tasks of the predicted bounding box. The solid wood panels selected in this paper are Chinese fir and pine. The defects include live knots, dead knots, decay, mildew, cracks, and pinholes. A total of more than 5000 samples were collected, and the data set was expanded to 100,000 through data enhancement methods. After using the improved SSD model, the average detection accuracy of the defects we obtained was 89.7%, and the average detection time was 90 ms. Both the detection accuracy and the detection speed were improved. |
| first_indexed | 2024-03-10T06:32:41Z |
| format | Article |
| id | doaj.art-5a38141442ef4af3b8407d0444019b37 |
| institution | Directory Open Access Journal |
| issn | 1999-4907 |
| language | English |
| last_indexed | 2024-03-10T06:32:41Z |
| publishDate | 2021-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Forests |
| spelling | doaj.art-5a38141442ef4af3b8407d0444019b372023-11-22T18:19:25ZengMDPI AGForests1999-49072021-10-011210141910.3390/f12101419Surface Detection of Solid Wood Defects Based on SSD Improved with ResNetYutu Yang0Honghong Wang1Dong Jiang2Zhongkang Hu3College of Mechanical and Electrical Engineering, Nanjing Forestry University, Nanjing 210037, ChinaCollege of Mechanical and Electrical Engineering, Nanjing Forestry University, Nanjing 210037, ChinaCollege of Mechanical and Electrical Engineering, Nanjing Forestry University, Nanjing 210037, ChinaNanjing Fujitsu Nanda Software Technology Co., Ltd., Nanjing 210012, ChinaDue to the lack of forest resources in China and the low detection efficiency of wood surface defects, the output of solid wood panels is not high. Therefore, this paper proposes a method for detecting surface defects of solid wood panels based on a Single Shot MultiBox Detector algorithm (SSD) to detect typical wood surface defects. The wood panel images are acquired by an independently designed image acquisition system. The SSD model included the first five layers of the VGG16 network, the SSD feature mapping layer, the feature detection layer, and the Non-Maximum Suppression (NMS) module. We used TensorFlow to train the network and further improved it on the basis of the SSD network structure. As the basic network part of the improved SSD model, the deep residual network (ResNet) replaced the VGG network part of the original SSD network to optimize the input features of the regression and classification tasks of the predicted bounding box. The solid wood panels selected in this paper are Chinese fir and pine. The defects include live knots, dead knots, decay, mildew, cracks, and pinholes. A total of more than 5000 samples were collected, and the data set was expanded to 100,000 through data enhancement methods. After using the improved SSD model, the average detection accuracy of the defects we obtained was 89.7%, and the average detection time was 90 ms. Both the detection accuracy and the detection speed were improved.https://www.mdpi.com/1999-4907/12/10/1419solid wood panelmachine visiondefects recognitionSSD |
| spellingShingle | Yutu Yang Honghong Wang Dong Jiang Zhongkang Hu Surface Detection of Solid Wood Defects Based on SSD Improved with ResNet Forests solid wood panel machine vision defects recognition SSD |
| title | Surface Detection of Solid Wood Defects Based on SSD Improved with ResNet |
| title_full | Surface Detection of Solid Wood Defects Based on SSD Improved with ResNet |
| title_fullStr | Surface Detection of Solid Wood Defects Based on SSD Improved with ResNet |
| title_full_unstemmed | Surface Detection of Solid Wood Defects Based on SSD Improved with ResNet |
| title_short | Surface Detection of Solid Wood Defects Based on SSD Improved with ResNet |
| title_sort | surface detection of solid wood defects based on ssd improved with resnet |
| topic | solid wood panel machine vision defects recognition SSD |
| url | https://www.mdpi.com/1999-4907/12/10/1419 |
| work_keys_str_mv | AT yutuyang surfacedetectionofsolidwooddefectsbasedonssdimprovedwithresnet AT honghongwang surfacedetectionofsolidwooddefectsbasedonssdimprovedwithresnet AT dongjiang surfacedetectionofsolidwooddefectsbasedonssdimprovedwithresnet AT zhongkanghu surfacedetectionofsolidwooddefectsbasedonssdimprovedwithresnet |