针对机器人视觉目标图像信噪比低、背景噪声干扰大的特点,采用马尔科夫随机场(Markov Random Field,MRF)模型的平滑去噪方法对图像进行预处理。在此基础上,采用K-means聚类算法对图像进行聚类,将具有不同特征的目标区域分类,为进一步实...针对机器人视觉目标图像信噪比低、背景噪声干扰大的特点,采用马尔科夫随机场(Markov Random Field,MRF)模型的平滑去噪方法对图像进行预处理。在此基础上,采用K-means聚类算法对图像进行聚类,将具有不同特征的目标区域分类,为进一步实现目标识别和跟踪提供基础。同时,为进一步克服移动机器人导航过程中视觉处理速度慢的缺陷,对图像进行分块划分,提取每个图像块的均值、方差和最大值作为特征值,从而提高算法的处理速度。展开更多
Clustered architecture is selected for high level synthesis,and a simultaneous partitioning and scheduling algorithm are proposed.Compared with traditional methods,circuit performance can be improved.Experiments show ...Clustered architecture is selected for high level synthesis,and a simultaneous partitioning and scheduling algorithm are proposed.Compared with traditional methods,circuit performance can be improved.Experiments show the efficiency of the method.展开更多
In order to exploit the capability of the peak-to-average power ratio(PAPR)reduction afforded by the partial transmit sequences (PTS)approach in orthogonal frequency division multiplexing(OFDM)systems, subblock ...In order to exploit the capability of the peak-to-average power ratio(PAPR)reduction afforded by the partial transmit sequences (PTS)approach in orthogonal frequency division multiplexing(OFDM)systems, subblock partition schemes for the PTS approach are studied. The motivation is to establish the relationship between the subblock partition and the capability of PAPR reduction through the periodic autocorrelation functions (ACFs)of partial transmit sequences and the periodic cross-correlation functions(CCFs)of signal candidates.Let Q represent the variation of the square magnitudes of ACFs.It is found that the lower the Q-value is, the better PAPR performance can be achieved, which is introduced as a design criterion for subblock partition.Based on this criterion, four common partition methods are compared and an efficient partition strategy is proposed. It is shown that structured partition schemes with low computational complexity have a large Q-value, leading to a poor PAPR performance.The new strategy can be regarded as a trade-off between PAPR performance and computational complexity.The simulation results show that the strategy can achieve an optimal performance with a relatively low complexity and, moreover,does not increase the amount of side information.展开更多
文摘针对机器人视觉目标图像信噪比低、背景噪声干扰大的特点,采用马尔科夫随机场(Markov Random Field,MRF)模型的平滑去噪方法对图像进行预处理。在此基础上,采用K-means聚类算法对图像进行聚类,将具有不同特征的目标区域分类,为进一步实现目标识别和跟踪提供基础。同时,为进一步克服移动机器人导航过程中视觉处理速度慢的缺陷,对图像进行分块划分,提取每个图像块的均值、方差和最大值作为特征值,从而提高算法的处理速度。
文摘Clustered architecture is selected for high level synthesis,and a simultaneous partitioning and scheduling algorithm are proposed.Compared with traditional methods,circuit performance can be improved.Experiments show the efficiency of the method.
文摘In order to exploit the capability of the peak-to-average power ratio(PAPR)reduction afforded by the partial transmit sequences (PTS)approach in orthogonal frequency division multiplexing(OFDM)systems, subblock partition schemes for the PTS approach are studied. The motivation is to establish the relationship between the subblock partition and the capability of PAPR reduction through the periodic autocorrelation functions (ACFs)of partial transmit sequences and the periodic cross-correlation functions(CCFs)of signal candidates.Let Q represent the variation of the square magnitudes of ACFs.It is found that the lower the Q-value is, the better PAPR performance can be achieved, which is introduced as a design criterion for subblock partition.Based on this criterion, four common partition methods are compared and an efficient partition strategy is proposed. It is shown that structured partition schemes with low computational complexity have a large Q-value, leading to a poor PAPR performance.The new strategy can be regarded as a trade-off between PAPR performance and computational complexity.The simulation results show that the strategy can achieve an optimal performance with a relatively low complexity and, moreover,does not increase the amount of side information.