研究目的
Investigating the effectiveness of two new thresholding methods for automated defect detection on titanium-coated aluminum surfaces.
研究成果
The proposed contrast-adjusted methods were able to correctly detect the highest number of predetermined defects in different sets of image resolution and performed better than MET and Otsu’s method. However, the lighting and shadow effects were found to influence the accuracy of defect detection.
研究不足
The lighting and shadow effects were found to influence the accuracy of defect detection.
1：Experimental Design and Method Selection:
The study proposed two new thresholding methods and compared them with four existing techniques.
2：Sample Selection and Data Sources:
Images of titanium-coated aluminum disk specimens were captured at high resolutions of 700, 900, and 1000 dpi.
3：List of Experimental Equipment and Materials:
A line sensor camera of high-spatial accuracy and color fidelity was used for image acquisition.
4：Experimental Procedures and Operational Workflow:
The images were processed using the proposed and existing thresholding methods, and the results were compared in terms of accuracy and speed.
5：Data Analysis Methods:
The performance of the thresholding methods was evaluated based on the number of defects detected and the time taken for detection.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
