.png){kind=icon}
.png){kind=icon}
Return
A visually compensated ultrasonic imaging method without position encoder
.png){kind=icon}
PRE
AI
DOI：10.1016/j.measurement.2025.116752.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the limited performance of traditional position encoders in terms of hardware cost and confined space, the employment of visual surveillance to obtain position information can effectively enhance the freedom of ultrasound detection. The YOLOv5s model is improved using the CBAM attention mechanism and the NWD loss function to achieve stable recognition of ultrasonic probes with a high detection accuracy of 94 % while maintaining a real-time detection rate of 21 fps. The acquired coordinate information is then fused with the postprocessed ultrasonic signal to form an ultrasound signal with coordinate information, which is used to visualize the internal condition of the material in the form of point cloud imaging. Finally, by measuring the 3D information of the defect point cloud and comparing it with the actual size, the depth error is below 5 % and the defect area error is below 14 %, which proves the effectiveness of the proposed method.
Keywords:
Ultrasonic
Nondestructive testing
CBAM
NWD
Damage imaging
Journal
IF:
5.6
Papers:
1.9W
Citations:
5.4W