基于AI技术的视频边缘检测算法研究.pdfVIP

控制理论与应用 《自动化技术与应用》2023年第42卷第3期 Control Theory and Applications DOI: 10.20033/j 1003-7241 (2023) 03-0045-05 基于Al技术的视频边缘检测算法研究舟 杜泽新 ，张孙蓉 ，崔珂伟 (贵州乌江水电开发有限责任公司水电站远程集控中心，贵州贵阳550002) 摘 要:为有效去除视频图像内部噪声，提升视频检测的边缘连续性 。利用块和低秩张量恢复滤除视频噪声;输入去噪后视频，求解 视频图像梯度与二阶导数;抑制梯度视频图像，获取梯度边缘;依据二阶导数检测视频图像边缘,获取视频图像的二阶导数边 缘;融合梯度与二阶导数边缘，输出最终视频边缘检测结果 。实验证明算法可有效去除视频噪声，提升视频清晰度;边缘检测 连续性较优;在不同噪声组合水平下，具备较优视频边缘检测性能。 关键词:AI技术;视频边缘检测算法；低秩张量；支持向量机；二阶导数 中图分类号:TP391.41 文献标识码:A 文章编号:1003-7241(2023)03-0045-05 Research on Video Edge Detection Algorithm Based on Al Tech no logy D Ze-xin, ZHANG Sun-rong, CUI Ke-wei (Remote Centralized Control Center of Hydropower Station of Guizhou Wujiang Hydropower Development Co., Ltd., Guiyang 550002 China) Abstract: In order to effectively remove the internal noise of video image and improve the edge continuity of video detection Block and low rank tensor recovery are used to filter video noise; Input the denoised video and solve the video image gradient and second derivative; Suppress the gradient video image and obtain the gradient edge; Detecting the edge of the video image according to the second derivative to obtain the second derivative edge of the video image; The gradient and second derivative edges are fused to output the final video edge detection results Experiments show that the algorithm can effectively remove video noise and im­ prove video definition, the continuity of edge detection is better It has better video edge detection perfonnance under different noise combination levels