论文部分内容阅读
针对传统的完全基于灰度阀值分割算法的转炉出钢下渣检测方法在遇到出钢过程中的意外情况时,检测准确率下降的缺点,提出了一种改进的红外转炉出钢下渣检测方法。该方法首先利用分水岭分割的方法将转炉出钢图像分割成几个区域,再利用形态学方法对分割出的区域进行矩形拟合以获取准确的钢流图像区域,最后通过前后共4帧图像信息的加权来消除出钢过程中的异常情况对于检测准确性的影响。现场试验结果表明,所提出的改进方法对于提高红外转炉出钢下渣检测系统报警准确率,减小下渣到钢包的渣厚度以及提高系统稳定性方面效果显著。
Aimed at the shortcomings of the traditional slag detection method based on gray threshold segmentation algorithm in the converter tapping slag detection in case of accidental tapping during the tapping process, an improved infrared converter slag tapping Detection method. Firstly, the method of watershed segmentation is used to segment the tapping of the converter into several regions, and the morphological method is used to fit the segmented regions to obtain the accurate steel flow image region. Finally, a total of four frames of image information Of the weight to eliminate tapping during the abnormal situation on the accuracy of detection. The field test results show that the proposed method is effective in improving the alarm accuracy of slag detection system at the tapping point of infrared converter, reducing the slag thickness from ladle to ladle, and improving system stability.