A new rnultiscale edge detection method is presented, which is based on an effective edge measure. The effective edge measure, used to adaptively adjust the scales of wavelet transform, is defined using the novel feat...A new rnultiscale edge detection method is presented, which is based on an effective edge measure. The effective edge measure, used to adaptively adjust the scales of wavelet transform, is defined using the novel features of image edge obtained from human being vision characteristics. Finally, two experiments show that the proposed algorithm appears to work well.展开更多
三维目标检测是实现无人驾驶必不可少的技术,但很多三维检测算法采用的分割算法并不能很好地提取局部特征,导致检测精度不理想。为了改善局部特征缺失的情况,提出一种基于边缘卷积的三维目标识别算法。本算法以激光点云和RGB(red,green,...三维目标检测是实现无人驾驶必不可少的技术,但很多三维检测算法采用的分割算法并不能很好地提取局部特征,导致检测精度不理想。为了改善局部特征缺失的情况,提出一种基于边缘卷积的三维目标识别算法。本算法以激光点云和RGB(red,green,blue)图像作为输入,基于二维候选区域中的像素过滤激光点云生成视锥点云,以此提高检测速度。同时,在分割算法中,在点云的局部特征图的基础上计算目标点和相邻点之间的欧氏距离,并将其作为边缘特征赋予目标点和相邻点。此外,在卷积神经网络提取特征的过程中,每次卷积完成后都会在新的局部特征图上重新计算三维点之间的欧氏距离,为三维点构造新的边缘特征。这使得边缘特征能随着卷积神经网络的计算扩散到整个点云,从而提高局部特征的提取效果。本算法在KITTI(Karlsruhe Institute of Technology and Toyota Technological Institute)的三维点云数据集上进行验证,分割精度达到92.82%,相比于F-PointNet提高了2.30百分点;对不同目标的检测精度也有所提高,车辆、自行车、行人的检测精度分别达到了85.77%,76.09%,53.08%。试验结果证明了本算法的可行性。本算法可应用于无人驾驶汽车,实现车辆、行人和自行车的定位与检测。展开更多
文摘A new rnultiscale edge detection method is presented, which is based on an effective edge measure. The effective edge measure, used to adaptively adjust the scales of wavelet transform, is defined using the novel features of image edge obtained from human being vision characteristics. Finally, two experiments show that the proposed algorithm appears to work well.
文摘三维目标检测是实现无人驾驶必不可少的技术,但很多三维检测算法采用的分割算法并不能很好地提取局部特征,导致检测精度不理想。为了改善局部特征缺失的情况,提出一种基于边缘卷积的三维目标识别算法。本算法以激光点云和RGB(red,green,blue)图像作为输入,基于二维候选区域中的像素过滤激光点云生成视锥点云,以此提高检测速度。同时,在分割算法中,在点云的局部特征图的基础上计算目标点和相邻点之间的欧氏距离,并将其作为边缘特征赋予目标点和相邻点。此外,在卷积神经网络提取特征的过程中,每次卷积完成后都会在新的局部特征图上重新计算三维点之间的欧氏距离,为三维点构造新的边缘特征。这使得边缘特征能随着卷积神经网络的计算扩散到整个点云,从而提高局部特征的提取效果。本算法在KITTI(Karlsruhe Institute of Technology and Toyota Technological Institute)的三维点云数据集上进行验证,分割精度达到92.82%,相比于F-PointNet提高了2.30百分点;对不同目标的检测精度也有所提高,车辆、自行车、行人的检测精度分别达到了85.77%,76.09%,53.08%。试验结果证明了本算法的可行性。本算法可应用于无人驾驶汽车,实现车辆、行人和自行车的定位与检测。
