改进的Gabor滤波器带钢表面缺陷显著性检测

Saliency defect detection in strip steel by improved Gabor filter

针对传统带钢表面缺陷检测方法在实际生产过程中检测精度低、实时性差的问题,提出一种基于复合差分进化的Gabor滤波器优化方法.首先,将采集到的图像进行预处理,获取高质量图像;然后,针对传统Gabor小波滤波器参数较多和算法实时性不高两大难题,提出了一种复合差分进化的Gabor滤波器优化方法,对参数和方向分别做了改进,较大提升了检测效率;最后,对显著性缺陷目标进行阈值分割,完成带钢表面缺陷检测.实验结果表明:该优化算法复杂度低、检测效率高,优化后的Gabor检测模型在速度上比传统Gabor检测模型快了约2.3倍,平均速度达到了91.8ms/帧.

Aiming at the defects of the traditional strip surface detection method in the actual production process cause many problems,such as low detection accuracy and real-time,an optimization method of Gabor filter was proposed based on composite differential evolution.Firstly,the high quality image was obtained after the collected images were preprocessed,and according to that the traditional Gabor wavelet filter parameters are more and the real-time performance of the algorithm is not high,an optimization method of Gabor filter was put forward based on composite differential evolution.The parameters and directions were improved respectively,and the detection efficiency was greatly improved.Finally,the strip surface defect detection experiment was completed after the significant defects of target segmentation.The results show that the algorithm has low complexity,high detection efficiency,and the detection of the optimized Gabor model is nearly 2.3times faster than the traditional Gabor detection model.The average speed of the image reaches 91.8ms/frame.