A Novel Image Processing Algorithm of Sand-Hole Extracted for Casting Parts

4th Annual International Workshop on Materials Science and Engineering

Lei, Ting^1,2 ; Xu, Jing^1

School of Mechatronics Engineering, Wuhan Donghu University, Wenhua Road, Wuhan, China^1

School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan, China^2

关键词: Casting parts;    Defect images;    High contrast;    High resolution;    Image processing algorithm;    Recognition methods;    Reference values;    Visual detection;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/381/1/012116/pdf
DOI  :  10.1088/1757-899X/381/1/012116

来源: IOP

【 摘 要 】

Focusing on the detecting problem for sand-hole of casting parts, a recognition method which based on morphologically reconstructing and differently subtracting algorithms was proposed. The image of casting whose surface in somewhere occupied with sand-hole was read and denoised by median filter. Then the image was histogram equalized, enhanced image obtained with high contrast and high resolution. Following process the defect image and the qualified image through denoising, enhancing and morphologically reconstructing, respectively. Achieving the difference information through subtracting the two images, the area with the largest grayscale value was the position of the sand hole. Finally, the area, the perimeter and the circularity of the sand hole are calculated by using the relevant function, so as to better identify the defects in the future. This paper has some reference value for the development of visual detection technology.