研究目的
To develop an adaptive algorithm for the analysis of thermal fatigue cracks on the surfaces of rollers of continuous casting machines that does not require adaptation to images of various types and individual choice of parameters.
研究成果
The developed algorithm with self-focusing guarantees the possibility of recognition of elongated curvilinear objects and, in particular, thermal fatigue cracks, in the images without special adaptation. The fuzzy set obtained in analyzing the image contains detailed information about the in-service defects of the surfaces of rollers of continuous casting machine recognized for different sets of values of the parameters.
研究不足
The proposed algorithm requires somewhat higher computational costs than the applied basic algorithm. The computational costs are much higher than the costs of the basic algorithm, necessitating the initial image to be made smaller to lower the costs.
1：Experimental Design and Method Selection:
The algorithm involves analyzing images for a sufficiently large subset of the sets of values of the parameters without requiring adaptation to images of various types and individual choice of parameters.
2：Sample Selection and Data Sources:
Images of the roller surface of a continuous casting machine with cracks for different levels of brightness and illuminance.
3：List of Experimental Equipment and Materials:
Not explicitly mentioned.
4：Experimental Procedures and Operational Workflow:
The algorithm includes the analysis of the images for sufficiently large subsets of the sets of values of the parameters of the basic algorithm. The result of this classification is regarded as a fuzzy set with the membership function of each element (pixel) equal to the number of sets of parameters with the help of which it is shown that this element belongs to the frame line of the crack.
5：Data Analysis Methods:
The function that gives the number of hits of a frame line by the point (x, y) for all investigated K sets of values is used for analysis.
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
