An industrial CT image segmentation algorithm based on non-parameter estimation

作者：Li, Jin; Wang, Lei^*; Bao, Peihua

来源：Journal of Computational Information Systems, 2010, 6(9): 3103-3109.

• 3-D space
• CAD models
• Conventional measurements
• Criterion functions
• Gray level image
• Image threshold
• Industrial components
• Industrial computed tomography
• Industrial CT
• Internal structure
• Laser scanner
• Low contrast
• Minimum cross entropy
• Non-parameter estimation
• Otsu method
• Parameter estimation method
• Parzen windows
• Processing Time
• Segmentation
• Threshold methods

Reverse engineering is a method for constructing a CAD model from a physical part. Compared to conventional measurement technologies such as 3D-laser scanner and so on, industrial computed tomography (ICT) is a technology which examines the internal structures of industrial components without destroying them. However, lots of artifacts and noise usually exists in the ICT images. Image segmentation for ICT images is one of the most important and fundamental tasks. The goal of our work is to decrease the noise and artifact in the ICT images by segmentation. Non-parameter estimation method such as the Parzen window technique is integrated to estimate the spatial probability distribution of gray-level image values, and a criterion function by the minimum cross entropy is used. In order to assess the effectiveness of the proposed threshold method, several real ICT images are used in the experiments. The results processed by the proposed method are compared with those yielded by three threshold techniques widely used in the literatures. The experimental results for ICT images demonstrate the success of the proposed image threshold method, as compared with the OTSU method and the moment-based method with histogram, which cannot segment the industrial CT images with low contrast ratio, much noise and artifact. For the future, the method will be extended to 3D space and improved to reduce the processing time.

• 出版日期2010
• 单位哈尔滨工程大学