Design and analysis of welding inspection robot

Zhang, Pengyu; Wang, Ji; Zhang, Feng; Xu, Penghui; Li, Leijun; Li, Baoming

doi:10.1038/s41598-022-27209-4

Cited by 6 publications

(7 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To solve the problem, the Hybrid Dilated Convolution (HDC) [69] criterion is used to design the size of the expansion rate. We designed the expansion rate to be (1,3,5) and the perception size to be 19 × 19, as shown in Figure 3c, to avoid the gridding effect. In following equations, we assume that the convolution kernel size of the dilation convolution is k and the stride is one, then the size of the receptive field of the expansion convolution of the i + first layer is r f i+1 :…”

Section: Dilated Convolution Aggregation Module (Dcam)mentioning

confidence: 99%

“…where I s is obtained by dilation convolution with different expansion rates (1,3,5) This paper uses the steel surface defect detection public dataset NEU-DET [72], which includes a total of 1800 images of six types of defects, 300 images of each defect type, and the size of the image is 200 × 200 pixels. We have renamed the six defect types for convenience: C (crazing), RS (rolled-in scales), I (inclusion), P (patches), PS (pitted surface), and S (scratches).…”

Section: Dilated Convolution and Spatial Attention Fusion Module (Dsm)mentioning

confidence: 99%

“…Steel accounts for more than 90% of all metals used in industrial production, because it is a material with high strength and ductility, and excellent manufacturability, at the lowest costs [ 1 ], ideal for machines, civil structures, transportation equipment, and endless list of tools [ 2 ]. As a metallic material [ 3 ], steel is widely used in manufacturing processes, such as brazing [ 4 , 5 , 6 ], laser welding [ 7 , 8 ], and additive manufacturing [ 9 ].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

DAssd-Net: A Lightweight Steel Surface Defect Detection Model Based on Multi-Branch Dilated Convolution Aggregation and Multi-Domain Perception Detection Head

et al. 2023

Sensors

Self Cite

View full text Add to dashboard Cite

During steel production, various defects often appear on the surface of the steel, such as cracks, pores, scars, and inclusions. These defects may seriously decrease steel quality or performance, so how to timely and accurately detect defects has great technical significance. This paper proposes a lightweight model based on multi-branch dilated convolution aggregation and multi-domain perception detection head, DAssd-Net, for steel surface defect detection. First, a multi-branch Dilated Convolution Aggregation Module (DCAM) is proposed as a feature learning structure for the feature augmentation networks. Second, to better capture spatial (location) information and to suppress channel redundancy, we propose a Dilated Convolution and Channel Attention Fusion Module (DCM) and Dilated Convolution and Spatial Attention Fusion Module (DSM) as feature enhancement modules for the regression and classification tasks in the detection head. Third, through experiments and heat map visualization analysis, we have used DAssd-Net to improve the receptive field of the model while paying attention to the target spatial location and redundant channel feature suppression. DAssd-Net is shown to achieve 81.97% mAP accuracy on the NEU-DET dataset, while the model size is only 18.7 MB. Compared with the latest YOLOv8 model, the mAP increased by 4.69%, and the model size was reduced by 23.9 MB, which has the advantage of being lightweight.

show abstract

Section: Dilated Convolution Aggregation Module (Dcam)mentioning

confidence: 99%

Section: Dilated Convolution and Spatial Attention Fusion Module (Dsm)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

DAssd-Net: A Lightweight Steel Surface Defect Detection Model Based on Multi-Branch Dilated Convolution Aggregation and Multi-Domain Perception Detection Head

et al. 2023

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…Given the insufficient weld seam training samples, a data augmentation method based on a deep semantic segmentation network is employed to expand the dataset, which is then used to train the lightweight YOLOv3 model, enabling the rapid localization of weld trajectories [ 26 ]. A weld seam detection and target recognition algorithm based on YOLOv5 is proposed for complex environments to ensure cost-effective computation, achieving a recognition accuracy exceeding 90% and improving operational efficiency by 12% [ 27 ]. A weld seam instance segmentation algorithm based on Mask R-CNN is proposed to address the influence of target detection on weld seam tracking accuracy, demonstrating a single-path planning time of 180 ms and an average precision of 67.6% [ 28 ].…”

Section: Introductionmentioning

confidence: 99%

Weld Seam Tracking and Detection Robot Based on Artificial Intelligence Technology

Wang

Huang

Yao

et al. 2023

Sensors

View full text Add to dashboard Cite

The regular detection of weld seams in large-scale special equipment is crucial for improving safety and efficiency, and this can be achieved effectively through the use of weld seam tracking and detection robots. In this study, a wall-climbing robot with integrated seam tracking and detection was designed, and the wall climbing function was realized via a permanent magnet array and a Mecanum wheel. The function of weld seam tracking and detection was realized using a DeepLabv3+ semantic segmentation model. Several optimizations were implemented to enhance the deployment of the DeepLabv3+ semantic segmentation model on embedded devices. Mobilenetv2 was used to replace the feature extraction network of the original model, and the convolutional block attention module attention mechanism was introduced into the encoder module. All traditional 3×3 convolutions were substituted with depthwise separable dilated convolutions. Subsequently, the welding path was fitted using the least squares method based on the segmentation results. The experimental results showed that the volume of the improved model was reduced by 92.9%, only being 21.8 Mb. The average precision reached 98.5%, surpassing the original model by 1.4%. The reasoning speed was accelerated to 21 frames/s, satisfying the real-time requirements of industrial detection. The detection robot successfully realizes the autonomous identification and tracking of weld seams. This study remarkably contributes to the development of automatic and intelligent weld seam detection technologies.

show abstract

“…In materials processing, the performance of a product is influenced by the chemical composition of the material [1][2][3][4][5][6][7][8], the processing method [9,10], the equipment, and the parameters [11]. Materials such as iron, copper, aluminum, and titanium have a wide range of applications in manufacturing, construction, electronics, and other industries.…”

Section: Introductionmentioning

confidence: 99%

Pre–Placed Metal Laser–Additive Manufacturing on Grade 5 Titanium Substrate: Layer Formation, Oxidation Resistance, and Wear Resistance

Lin,

Wang,

et al. 2023

Metals

Self Cite

View full text Add to dashboard Cite

The laser parameters and additive materials were investigated for the laser–additive manufacturing of titanium alloys. A pre–placed metal–strip–based method for 3D printing titanium components was proposed. Before laser processing, the metal strips were fixed to a Grade 5 titanium substrate using resistance spot welding. The effect of the processing conditions (laser power: 1.5–3.0 kW; scanning speed: 3–9 mm·s−1; one or three layers; 1, 6, or 12 passes; Grade 2 and Grade 5 as additives) on the microstructures, oxidation resistance, and wear resistance of the as–printed samples was investigated. The results showed that the microstructure consisted of α′ martensite, oxygen–stabilized α, transformed β that contains coarse and fine acicular α, titanium oxides, and carbides. Cracks were observed, particularly near the upper surface of the three–layer samples. The cracks were suggested to have formed due to cumulative residual stresses and the formation of oxygen–stabilized α and α′ martensite that might cause embrittlement. Both oxidation and wear tests were conducted to verify the improved performance. After 55 h of oxidation, the as–printed samples showed mass gains of 0.029–0.035 g·cm−2, which were smaller than those of the substrate (0.039 g·cm−2).

show abstract

Design and analysis of welding inspection robot

Cited by 6 publications

References 31 publications

DAssd-Net: A Lightweight Steel Surface Defect Detection Model Based on Multi-Branch Dilated Convolution Aggregation and Multi-Domain Perception Detection Head

DAssd-Net: A Lightweight Steel Surface Defect Detection Model Based on Multi-Branch Dilated Convolution Aggregation and Multi-Domain Perception Detection Head

Weld Seam Tracking and Detection Robot Based on Artificial Intelligence Technology

Pre–Placed Metal Laser–Additive Manufacturing on Grade 5 Titanium Substrate: Layer Formation, Oxidation Resistance, and Wear Resistance

Contact Info

Product

Resources

About