针对目前城市道路场景中行道树提取方法需要设置的参数较多以及树冠点云相互重叠难以精确分割的问题,文章采用一种行道树提取与单株木分割算法。首先通过布料滤波算法从原始点云中移除地面点,并利用半径滤波滤除离群点,去除地面点和噪...针对目前城市道路场景中行道树提取方法需要设置的参数较多以及树冠点云相互重叠难以精确分割的问题,文章采用一种行道树提取与单株木分割算法。首先通过布料滤波算法从原始点云中移除地面点,并利用半径滤波滤除离群点,去除地面点和噪声点对行道树提取的影响;然后通过增加PointNet++网络的点集抽象模块(set abstraction,SA)提高模型特征提取能力,使模型更适用于行道树点云的提取,并利用改进后的网络从原始点云中提取行道树点云;最后结合密度聚类算法(density-based spatial clustering of applications with noise,DBSCAN)与K-Means算法对相互重叠的行道树点云进行分割,得到单株木信息。为验证该方法的有效性,以北京永昌路道路数据集进行训练测试。结果表明:改进后模型的行道树点云平均提取精度和交并比(intersection over union,IoU)分别提高了9.2%和15.1%,达到了94.5%、0.916;单木分割平均精度达到了91.3%。展开更多
Currently,investigations of the color attributes of Foshan Woodblock New Year Paintings are limited to qualitative descriptions,devoid of objective quantitative data backing.Their color solutions fail to satisfy the d...Currently,investigations of the color attributes of Foshan Woodblock New Year Paintings are limited to qualitative descriptions,devoid of objective quantitative data backing.Their color solutions fail to satisfy the distribution and creative requirements of the digital era.This study intends to utilize K-means clustering to calculate and analyze the color properties of Foshan Woodblock New Year Paintings.Incorporating contemporary display standards,this study examines color palettes consistent with traditional New Year Paintings culture.Firstly,a dataset consisting of 127 photos categorized into 6 principal types of Foshan Woodblock New Year Paintings was assembled.Images were subjected to pre-processing,which included color correction and noise reduction.Then,following the conversion of data from RGB to HSV color space to extract pixel vectors,the K-means algorithm was implemented with a specified K value for clustering.Primary colors were derived from both the comprehensive information and specific groups to create color schemes.Ultimately,in light of contemporary digital display attributes,the S and V values of cluster centers were moderately modified while maintaining conventional color characteristics to augment visual impact in digital distribution.The clustering results showed that Foshan Woodblock New Year Paintings display attributes of high saturation and pronounced value contrast,with prominent colors primarily consisting of high-saturation reds.This resulted in one comprehensive color scheme and six categorized schemes.This study offers methodological assistance for the quantitative analysis of New Year painting colors,providing practical recommendations for the innovative distribution of traditional New Year Paintings in contemporary media,and is crucial for the preservation and transmission of intangible cultural heritage.展开更多
在室内可见光通信中符号间干扰和噪声会严重影响系统性能,K均值(K-means)均衡方法可以抑制光无线信道的影响,但其复杂度较高,且在聚类边界处易出现误判。提出了改进聚类中心点的K-means(Improved Center K-means,IC-Kmeans)算法,通过随...在室内可见光通信中符号间干扰和噪声会严重影响系统性能,K均值(K-means)均衡方法可以抑制光无线信道的影响,但其复杂度较高,且在聚类边界处易出现误判。提出了改进聚类中心点的K-means(Improved Center K-means,IC-Kmeans)算法,通过随机生成足够长的训练序列,然后将训练序列每一簇的均值作为K-means聚类中心,避免了传统K-means反复迭代寻找聚类中心。进一步,提出了基于神经网络的IC-Kmeans(Neural Network Based IC-Kmeans,NNIC-Kmeans)算法,使用反向传播神经网络将接收端二维数据映射至三维空间,以增加不同簇之间混合数据的距离,提高了分类准确性。蒙特卡罗误码率仿真表明,IC-Kmeans均衡和传统K-means算法的误码率性能相当,但可以显著降低复杂度,特别是在信噪比较小时。同时,在室内多径信道模型下,与IC-Kmeans和传统Kmeans均衡相比,NNIC-Kmeans均衡的光正交频分复用系统误码率性能最好。展开更多
文摘针对目前城市道路场景中行道树提取方法需要设置的参数较多以及树冠点云相互重叠难以精确分割的问题,文章采用一种行道树提取与单株木分割算法。首先通过布料滤波算法从原始点云中移除地面点,并利用半径滤波滤除离群点,去除地面点和噪声点对行道树提取的影响;然后通过增加PointNet++网络的点集抽象模块(set abstraction,SA)提高模型特征提取能力,使模型更适用于行道树点云的提取,并利用改进后的网络从原始点云中提取行道树点云;最后结合密度聚类算法(density-based spatial clustering of applications with noise,DBSCAN)与K-Means算法对相互重叠的行道树点云进行分割,得到单株木信息。为验证该方法的有效性,以北京永昌路道路数据集进行训练测试。结果表明:改进后模型的行道树点云平均提取精度和交并比(intersection over union,IoU)分别提高了9.2%和15.1%,达到了94.5%、0.916;单木分割平均精度达到了91.3%。
文摘Currently,investigations of the color attributes of Foshan Woodblock New Year Paintings are limited to qualitative descriptions,devoid of objective quantitative data backing.Their color solutions fail to satisfy the distribution and creative requirements of the digital era.This study intends to utilize K-means clustering to calculate and analyze the color properties of Foshan Woodblock New Year Paintings.Incorporating contemporary display standards,this study examines color palettes consistent with traditional New Year Paintings culture.Firstly,a dataset consisting of 127 photos categorized into 6 principal types of Foshan Woodblock New Year Paintings was assembled.Images were subjected to pre-processing,which included color correction and noise reduction.Then,following the conversion of data from RGB to HSV color space to extract pixel vectors,the K-means algorithm was implemented with a specified K value for clustering.Primary colors were derived from both the comprehensive information and specific groups to create color schemes.Ultimately,in light of contemporary digital display attributes,the S and V values of cluster centers were moderately modified while maintaining conventional color characteristics to augment visual impact in digital distribution.The clustering results showed that Foshan Woodblock New Year Paintings display attributes of high saturation and pronounced value contrast,with prominent colors primarily consisting of high-saturation reds.This resulted in one comprehensive color scheme and six categorized schemes.This study offers methodological assistance for the quantitative analysis of New Year painting colors,providing practical recommendations for the innovative distribution of traditional New Year Paintings in contemporary media,and is crucial for the preservation and transmission of intangible cultural heritage.
文摘在高压并联电抗器声纹信号监测系统中,长时海量无标签声纹的高维非平稳性导致特征提取困难、无监督聚类适应性差。由此提出了一种基于深度自适应K-means++算法(deep adaptive K-means++clustering algorithm,DAKCA)的750 kV电抗器声纹聚类方法。首先通过采用两阶段无监督策略微调的改进堆叠稀疏自编码器(stacked sparse autoencoder,SSAE),对快速傅里叶变换后的归一化频域数据提取电抗器原始声纹32维深度特征。进一步提出了依据最近邻聚类有效性指标(clustering validation index based on nearest neighbors,CVNN)的自适应K-means++聚类算法,构建了能自适应确定最优聚类个数的电抗器声纹聚类模型。最后通过西北地区某750 kV电抗器实测声纹数据集进行了验证。结果表明,DAKCA算法对无标签声纹数据在不同样本均衡程度下能够稳定提取32维深度特征,并实现最优聚类,为直接高效利用电抗器无标签声纹数据提供了参考。
文摘在室内可见光通信中符号间干扰和噪声会严重影响系统性能,K均值(K-means)均衡方法可以抑制光无线信道的影响,但其复杂度较高,且在聚类边界处易出现误判。提出了改进聚类中心点的K-means(Improved Center K-means,IC-Kmeans)算法,通过随机生成足够长的训练序列,然后将训练序列每一簇的均值作为K-means聚类中心,避免了传统K-means反复迭代寻找聚类中心。进一步,提出了基于神经网络的IC-Kmeans(Neural Network Based IC-Kmeans,NNIC-Kmeans)算法,使用反向传播神经网络将接收端二维数据映射至三维空间,以增加不同簇之间混合数据的距离,提高了分类准确性。蒙特卡罗误码率仿真表明,IC-Kmeans均衡和传统K-means算法的误码率性能相当,但可以显著降低复杂度,特别是在信噪比较小时。同时,在室内多径信道模型下,与IC-Kmeans和传统Kmeans均衡相比,NNIC-Kmeans均衡的光正交频分复用系统误码率性能最好。
