Study Of Steel Slag Detection Method Based On Machine Vision Flow Folder

Slag making inevitably brings about slag inclusion phenomenon. For this reason, it is necessary to detect the slag ratio of steel to control the purity of steel in real time. In this paper, a deep research and comparison was made on the advantages and disadvantages of the domestic and international slag detection methods, combined with the technological process of the domestic converter steelmaking then studied the slag detection method based on machine vision video processing.By analyzing and comparing the infrared radiance differences of slag and pure steel on the theory of blackbody radiation, considering the costs of hardware equipment as a whole, we chosed the infrared camera as the capture device and captured the steel stream video frames by writing programs in C++. By analyzing and comparing the processing effects of different image enhancement algorithms and image segmentation algorithms in processing typical steel stream pictures, two kinds of slag detection algorithm models were developed, one for processing the steel stream pictures with the1-D OTSU algorithm which was modified by adding the coefficient n1×n2or the precise2-D OTSU algorithm, the other for processing the steel stream pictures with the image enhancement algorithm firstly and then with the1-D OTSU algorithm which was modified by adding the coefficient n1×n2or the precise2-D OTSU algorithm secondly. The proposed algorithm models were used to process the steel stream pictures to obtain the segmentation threshold values which were used to accurately calculate the ratio of slag.The proposed method had been tested in the laboratory environment and the results showed it achieved the requirements of real-time, accuracy and stability.