光照不均下图像明暗区域关键目标增强仿真
|
摘要
对光照不均图像的关键区域进行目标增强,能够有效提高明暗不均图像的处理效率以及处理精度。对图像明暗反差较大区域的关键目标的增强,需要先计算聚类的目标函数,进而确定聚类中心和隶属矩阵,完成图像关键目标的高效增强C传统方法通过图像特征点在RGB颜色空间中的分布规律,利用统计直方图理论模型对关键目标进行初步分割,但忽略了对目标隶属矩阵的确定,导致目标增强效果不理想。提出基于K-means聚类的光照不均图像明暗区域关键目标增强方法,首先获取目标轮廓并恢复其对应像素点,在灰度共生矩阵构建的基础之上,分别对能量、熵、相关性以及对比度进行计算,使得反差较大特征提取效果更好;以以上步骤为依据,对聚类目标函数进行计算,并根据对聚类中心和隶属矩阵的确定,完成图像明暗反差较大区域关键目标的增强。实验结果表明,所提方法在目标增强方面均优于当前方法,同时提高了光照不均对境下图像处理的准确性。
This paper focuses on a method for enhancing key objectives in light and dark area of image with uneven illumination based on K-means clustering. Firstly, this method obtained the object contour and restored its corresponding pixels. On the basis of the gray co-occurrence matrix, our research calculated energy, entropy, correlatior and contrast, the feature extraction effect with sharp contrast could be better. According to above steps, this research calculated the clustering objective function. After determining the cluster centers and membership matrix, we completed the enhancement of key objectives in light and dark areas of image with uneven illumination. Simulation results prove that the proposed method is better than the current method in target enhancement, which improves the accuracy of image processing in environment of uneven illumination.
This paper focuses on a method for enhancing key objectives in light and dark area of image with uneven illumination based on K-means clustering. Firstly, this method obtained the object contour and restored its corresponding pixels. On the basis of the gray co-occurrence matrix, our research calculated energy, entropy, correlatior and contrast, the feature extraction effect with sharp contrast could be better. According to above steps, this research calculated the clustering objective function. After determining the cluster centers and membership matrix, we completed the enhancement of key objectives in light and dark areas of image with uneven illumination. Simulation results prove that the proposed method is better than the current method in target enhancement, which improves the accuracy of image processing in environment of uneven illumination.
引文
[1]刘昌振,等.基于多尺度分割的高分遥感图像变异函数纹理提取和分类[J].国土资源遥感,2015,27(4):47-53.
[2]谭春超,杨洁明.煤与矸石图像灰度信息和纹理特征的提取研究[J].工矿自动化,2017,43(4):27-31.
[3]肖进胜,等.基于轮廓模板和自学习的图像纹理增强超采样算法[J].自动化学报,2016,42(8):1248-1258.
[4]王威,等.基于DCT域纹理结构相似度的模糊图像质量评价[J].计算机工程,2015,41(11):253-256.
[5]罗三定,彭琼,李婷.瓷砖图像的纹理特征分类研究[J].计算机工程与应用,2016,52(8):196-200.
[6]王雪,张为民.基于纹理和邻域信息的肤色分割算法[J].吉林大学学报:理学版,2017,55(4):909-914.
[7]丁海勇,郭瑞瑞,罗海滨.顾及纹理信息的遥感图像NSCT域自适应阈值去噪[J].遥感技术与应用,2017,32(3):435-442.
[8]李彦飞.关于弱纹理人脸图像匹配方法的改进研究[J].计算机仿真,2016,33(7):331-334.
[9]肖淑苹,贺毅岳.一种改进的EMD图像信号去噪算法[J].现代电子技术,2016,39(16):91-93.
[10]乔双,李健,孙佳宁.射线图像的快速纹理提取算法[J].原子能科学技术,2015,49(5):939-943.
[2]谭春超,杨洁明.煤与矸石图像灰度信息和纹理特征的提取研究[J].工矿自动化,2017,43(4):27-31.
[3]肖进胜,等.基于轮廓模板和自学习的图像纹理增强超采样算法[J].自动化学报,2016,42(8):1248-1258.
[4]王威,等.基于DCT域纹理结构相似度的模糊图像质量评价[J].计算机工程,2015,41(11):253-256.
[5]罗三定,彭琼,李婷.瓷砖图像的纹理特征分类研究[J].计算机工程与应用,2016,52(8):196-200.
[6]王雪,张为民.基于纹理和邻域信息的肤色分割算法[J].吉林大学学报:理学版,2017,55(4):909-914.
[7]丁海勇,郭瑞瑞,罗海滨.顾及纹理信息的遥感图像NSCT域自适应阈值去噪[J].遥感技术与应用,2017,32(3):435-442.
[8]李彦飞.关于弱纹理人脸图像匹配方法的改进研究[J].计算机仿真,2016,33(7):331-334.
[9]肖淑苹,贺毅岳.一种改进的EMD图像信号去噪算法[J].现代电子技术,2016,39(16):91-93.
[10]乔双,李健,孙佳宁.射线图像的快速纹理提取算法[J].原子能科学技术,2015,49(5):939-943.