脚部纹路演化的糖尿病识别方法研究

研究脚部纹路演化的糖尿病识别准确度优化问题。脚部纹路的变化可以反映糖尿病患者的早期症状,但是由于脚部纹路较为复杂,趾与趾之间,趾腹间的纹路不明显,给脚部纹路图像提取与分割带来了困难。传统的图像分割过程由于趾与趾之间,趾腹间的纹路不清晰漏掉细节信息,造成分割效果不好,判断结果不准。为了避免上述缺陷,提出了一种基于脚部纹路演化图像分割的早期糖尿病自动识别方法。利用奇异值分解方法,对采集的脚部纹路演化图像进行特征提取,为早期糖尿病识别提供依据。建立GCA演化模型,获取脚部纹路分布信息,能够实现脚部纹路演化图像的分割处理,从而完成早期糖尿病的自动识别。实验结果表明,应用改进算法进行早期糖尿病自动识别,能够及时发现糖尿病的前兆,从而对糖尿病进行早期治疗。

缺失指纹图像识别技术研究与仿真

研究指纹图像缺失的识别问题,当提取的指纹存在破损、缺失等不完整的问题的时候,传统的指纹识别算法都是对完整的指纹图像进行直接的操作,使算法很难准确的识别。针对上述问题,提出了指纹图像曲线变化的图像修补识别算法。通过分析指纹曲线轮廓的变化,计算出指纹曲线的演化规律,再通过离散迭代方法,完成指纹缺失图像的修补,进行识别。实验表明,方法能够有效的修补缺失的指纹图像部分,提高了后期基于指纹图像识别的准确度,取得了比较好的效果。

Effects of wheelset vibration on initiation and evolution of rail short-pitch corrugation

The initiation and evolution of short-pitch corrugation in Beijing metro line 4 was studied from the viewpoint of wheelset vibration.A three-dimensional elastic model was set up.Numerical simulations were undertaken with this model to analyze the corrugation by the wheelset vertical vibration and torsional vibration.Based on numerical results,the relation between rail corrugation and wheelset vibration,and the relation between the position of electromotor and wheelset vibration were indicated.It is found that avoiding the wheelset-rail resonance is one method of controlling the rail short-pitch corrugation and solving the vibration and noise problem in metro lines.