| 改进Laplace的无人机图像边缘检测算法研究 |
| |
| 引用本文： | 陈思吉,王晓红,李运川.改进Laplace的无人机图像边缘检测算法研究[J].测绘工程,2021,30(2):36-44. |
| |
| 作者姓名： | 陈思吉 王晓红 李运川 |
| |
| 作者单位： | 贵州大学 矿业学院,贵州 贵阳 550025;贵州大学 林学院,贵州 贵阳 550025;贵州大学 矿业学院,贵州 贵阳 550025 |
| |
| 基金项目： | 贵州省科技计划课题;贵州省自然科学基金资助项目;国家自然科学基金资助项目 |
| |
| 摘 要： | 当图像灰度发生急剧变化时,常规算法存在“阈值检测不敏感”的问题。针对这一问题,借鉴RGB三相分解变换的思路,基于Laplace算法基本原理,文中提出BRGB-ALaplace算法。该算法先进行图像增强,并用高斯滤波器平滑图像和抑制噪声;再对RGB 3个相位分量方向进行拉普拉斯模板锐化拉伸,对这3个分量进行相位重组,实现基于Laplace边缘检测,进而对Laplace算法进行改进;最后,选取第三次全国土地调查中无人机外业举证脱敏图像进行实验,将提出的改进算法与常规算法进行检测效果比较,并开展有无噪声两种情况下的实验对比分析。实验结果评估显示,文中提出的BRGB-ALaplace算法边缘检测效果更好,所获得的峰值信噪比(PSNR)高于常规算法,且均方根误差(MSE)更小,算法也具有更好的鲁棒性。试验结果表明,提出的BRGB-ALaplace算法在实际生产应用中能够有效提高噪声图像边缘检测的效果,且因其检测结果中含有少量纹理信息,能够辅助地物识别,减少地物误判。
|
| 关 键 词： | 边缘检测 拉普拉斯算法 峰值信噪比 鲁棒性 无人机图像 均方根误差 |
Research on improved Laplace UAV image edge detection algorithm |
| |
| CHEN Siji,WANG Xiaohong,LI Yunchuan.Research on improved Laplace UAV image edge detection algorithm[J].Engineering of Surveying and Mapping,2021,30(2):36-44. |
| |
| Authors: | CHEN Siji WANG Xiaohong LI Yunchuan |
| |
| Institution: | (College of Mining, Guizhou University, Guiyang 550025, China;College of Forestry, Guizhou University, Guiyang 550025,China) |
| |
| Abstract: | Edge detection is the basic link of digital image processing.When the image gray level changes drastically,the conventional algorithm has the problem of“insensitive threshold detection”.In view of this problem,drawing on the idea of RGB three-phase decomposition transformation,based on the basic principle of Laplace algorithm,this paper proposes a BRGB-ALaplace(Advanced Laplace-based algorithm based on RGB)algorithm.The algorithm first enhances the image,smoothes the image with a Gaussian filter,suppresses noise,performs Laplacian tem-plate sharpening and stretching on the three RGB phase component directions,performs phase reorganization on these three components to achieve Laplace edge Detection,and then improve the Laplace algorithm.Finally,the paper selects the demonstrative desensitization image of the drone field proof in the third national land survey,compares the proposed improved algorithm with the conventional algorithm to detect the effect,and conducts a comparative analysis of the experiment with and without noise The evaluation of the experimental result shows that the edge detection effect of the BRGB-ALaplace algorithm proposed in this paper is better,the peak signal-to-noise ratio(PSNR)obtained is higher than that of the conventional algorithm,the root mean square error(MSE)is smaller,and the algorithm is better Robustness.The experimental result shows that the proposed BRGB-ALaplace algorithm can effectively improve the edge detection effect of noisy images in actual production applications,and it can assist in feature recognition and reduce misjudgment of features because its detection results contain a small amount of texture information. |
| |
| Keywords: | edge detection Laplace algorithm peak signal-to-noise ratio robustness drone image root mean square error |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
