管道机器人采集图像缺陷检测方法研究

Research on defect detection method of pipeline robot image acquisition

针对管道机器人采集管道缺陷图像边缘提取准确性较低的问题,提出一种基于自适应的管道图像缺陷检测方法。首先,采用引导滤波对单尺度Retinex算法进行改进,实现自适应增强管道图像,并通过双边滤波改进Canny算法进,实现管道缺陷的有效提取与检测;然后,结合自适应图像增强算法与缺陷检测算法,详细设计了管道机器人缺陷检测算法流程;最后,通过在饱和蒸汽Q235A(φ300×8 mm)管道进行缺陷检测,对提出算法进行了验证。结果表明,该研究算法可自适应调节图像亮度达到亮度均衡,并良好地保留图像纹理细节,对管道缺陷的检测识别准确率可达到97%,相较于对比算法,直方图均衡化算法、SLVM算法、同态滤波算法,该研究算法的标准差平均降低了71.2%,平均梯度提升了15.1%,峰值信噪比提升了9.6%。

Aiming at the problem that the accuracy of edge extraction of pipeline defect image collected by pipeline robot is low,an adaptive defect detection method is proposed.Firstly,guided filter is used to improve the single scale Retinex algorithm to realize the adaptive enhancement of pipeline image,and bilateral filter is used to improve Canny algorithm to realize the defect extraction and detection.Then,combined with adaptive image enhancement algorithm and defect detection algorithm,the defect detection algorithm flow of pipeline robot is designed in detail;Finally,through the saturated steam Q235A(φ300×8 mm)pipeline,and the proposed algorithm is verified.The results show that the algorithm can adaptively adjust the image brightness to achieve brightness balance,and retain the image texture details.The detection and recognition accuracy of pipeline defects can reach 97%.Compared with the contrast algorithm,histogram equalization algorithm,SLVM algorithm and homomorphic filtering algorithm,the standard deviation of the algorithm is reduced by 71.2%,and the average gradient is increased by 15.1%,The PSNR is improved by 9.6%.