针对工业现场中复杂背景下镁熔液弱小目标实时检测的难题, 从图像识别的角度提出了基于 Prewitt算子的自适应背景预测算法。该算法首先使用 Prewitt算子对原始图像进行处理从而计算出图像最大灰度差, 其次根据最大灰度差与每个像素点的灰度差的差异选择背景预测模型进行处理得到背景预测图像, 然后用原始图像减去背景预测图像得到残差图像, 接着对残差图像作帧差运算以及阈值分割运算得到二值图像, 最后使用形态学运算获取最终的目标, 并将该算法与最小一乘法的检测性能进行对比。 Matlab仿真结果表明, 该算法不仅可检测到弱小目标, 并且检测到的目标点面积增大了 60%, 检测时间减少了 96.92%, 为图像处理技术应用于工业现场实时检测镁熔液中弱小目标奠定了基础。

Aiming at the problem of dim and weak target real-time detection in magnesium melt under a complex background, an adaptive background prediction algorithm based on the Prewitt operator is proposed from the perspective of image recognition. First, the Prewitt algorithm is used to process the original image of the molten magnesium to calculate the maximum gray difference of the image. Second, according to the difference between the maximum gray difference and the gray difference of each pixel, the background prediction model is selected to get the background image. Third, the original image is subtracted from the background image to obtain the residual image. Then, the two-value image is obtained by the frame difference operation and threshold segmentation. Finally, morphological operation is used to obtain the final target. The detection performance of the algorithm is compared with that of the least absolute deviation. Matlab simulation results show that this algorithm can detect the dim and weak target; also, the target area increased by 60% and the detection time was reduced by approximately 96.92%. This work lays a foundation for the application of image-processing technology in the field of real-time measurement of hydrogen content in magnesium melt.