为解决双目视觉三维重建深度图边缘不连续的问题,提出基于加权最小二乘(weighted least squares,WLS)滤波的深度图优化,经双目标定、畸变矫正、立体校正、立体匹配建立三维深度图,加入WLS滤波,通过调整正则项更改约束条件,对梯度较大的...为解决双目视觉三维重建深度图边缘不连续的问题,提出基于加权最小二乘(weighted least squares,WLS)滤波的深度图优化,经双目标定、畸变矫正、立体校正、立体匹配建立三维深度图,加入WLS滤波,通过调整正则项更改约束条件,对梯度较大的区域减少约束,保留图像边缘,对梯度较小的区域平滑处理,去除噪声,采用峰值信噪比、结构相似性指数、平均绝对误差3个参数评价图像质量。评价结果表明:与半全局匹配算法相比,此算法的峰值信噪比增大1.849 dB,图像失真更少,质量更高;结构相似性指数增大0.4151,与原图结构相似性更强;平均绝对误差减小21.5422,还原度更高。重建的深度图视觉效果更好,改善立体匹配不连续的问题,减小匹配误差,使视差图质量更高。展开更多
三维定位是实现采茶机器人精采名优茶的关键技术,对保证机器人采摘茶叶高品质和高产量具有重要的意义,传统的SGBM(Semi-Global Block Matching)算法存在匹配效果差,还原效果不高等问题。本文提出SGBM算法与WLS(Weighted Least Squares)...三维定位是实现采茶机器人精采名优茶的关键技术,对保证机器人采摘茶叶高品质和高产量具有重要的意义,传统的SGBM(Semi-Global Block Matching)算法存在匹配效果差,还原效果不高等问题。本文提出SGBM算法与WLS(Weighted Least Squares)的融合算法,使得茶叶嫩芽深视图轮廓更清晰、前后景分层更明显、还原度更高,实际定位效果更精准。实验表明:采用SGBM与WLS融合算法能够将定位误差控制在1 mm左右,约是同等条件下其他传统融合算法精确度的7倍,提高了机器人采摘茶叶时定位的工作效率,为后续实现采茶机器人智能化提供一定帮助。展开更多
In this paper,the design,construction,simulation,and performance of a position-sensitive plastic scintillator detector are presented.The readout of the detector uses wavelength-shifting fibers coupled with multi-anode...In this paper,the design,construction,simulation,and performance of a position-sensitive plastic scintillator detector are presented.The readout of the detector uses wavelength-shifting fibers coupled with multi-anode photomultipliers(PMTs) for the x-and y-dimensions.After calibrating the multi-anode PMTs,a two-dimensional projection image of the square scintillator telescope hit by cosmic muons is demonstrated.By performing a cosmic test with the Micromegas telescope,the position resolution of the detector was determined to reach approximately8.6 mm,which is close to the value obtained in physical simulation.展开更多
文摘为解决双目视觉三维重建深度图边缘不连续的问题,提出基于加权最小二乘(weighted least squares,WLS)滤波的深度图优化,经双目标定、畸变矫正、立体校正、立体匹配建立三维深度图,加入WLS滤波,通过调整正则项更改约束条件,对梯度较大的区域减少约束,保留图像边缘,对梯度较小的区域平滑处理,去除噪声,采用峰值信噪比、结构相似性指数、平均绝对误差3个参数评价图像质量。评价结果表明:与半全局匹配算法相比,此算法的峰值信噪比增大1.849 dB,图像失真更少,质量更高;结构相似性指数增大0.4151,与原图结构相似性更强;平均绝对误差减小21.5422,还原度更高。重建的深度图视觉效果更好,改善立体匹配不连续的问题,减小匹配误差,使视差图质量更高。
基金supported by the Strategic Priority Research Program on Space Science of Chinese Academy of Sciences(No.XDA04070900)
文摘In this paper,the design,construction,simulation,and performance of a position-sensitive plastic scintillator detector are presented.The readout of the detector uses wavelength-shifting fibers coupled with multi-anode photomultipliers(PMTs) for the x-and y-dimensions.After calibrating the multi-anode PMTs,a two-dimensional projection image of the square scintillator telescope hit by cosmic muons is demonstrated.By performing a cosmic test with the Micromegas telescope,the position resolution of the detector was determined to reach approximately8.6 mm,which is close to the value obtained in physical simulation.
