Blood vessels in ophthalmoscope images play an important role in diagnosis of some serious pathology on retinal images. Hence, accurate extraction of vessels is becoming a main topic of this research area. In this pap...Blood vessels in ophthalmoscope images play an important role in diagnosis of some serious pathology on retinal images. Hence, accurate extraction of vessels is becoming a main topic of this research area. In this paper, a new hybrid approach called the (Genetic algorithm and vertex chain code) for blood vessel detection. And this method uses geometrical parameters of retinal vascular tree for diagnosing of hypertension and identified retinal exudates automatically from color retinal images. The skeletons of the segmented trees are produced by thinning. Three types of landmarks in the skeleton must be detected: terminal points, bifurcation and crossing points, these points are labeled and stored as a chain code. Results of the proposed system can achieve a diagnostic accuracy with 96.0% sensitivity and 98.4% specificity for the identification of images containing any evidence of retinopathy.展开更多
合理的集料级配能有效提升道路性能并延长道路使用寿命,针对拌合站现场需要快速进行集料级配检测的诉求,设计基于双参量MCMC算法的集料级配检测系统。针对单视角集料图像无法准确反映集料实际粒径的问题,该系统设置双相机进行图像采集;...合理的集料级配能有效提升道路性能并延长道路使用寿命,针对拌合站现场需要快速进行集料级配检测的诉求,设计基于双参量MCMC算法的集料级配检测系统。针对单视角集料图像无法准确反映集料实际粒径的问题,该系统设置双相机进行图像采集;并引入基于凹点匹配的图像分割算法,解决副相机采集过程中出现的颗粒堆叠问题;为尽可能准确地表征集料颗粒等效粒径,经实验验证选择等效椭圆短径均值和等效Feret短径均值分别表征不同粒径区间内集料颗粒;将双相机采集图像中提取到的集料颗粒信息作为双参量输入,利用双参量贝叶斯统计思想对集料实际粒径分布进行推断;在工程上引入马尔科夫链蒙特卡洛(Markov chain Monte Carlo,MCMC)算法计算双参量贝叶斯后验分布,并将其作为集料级配结果输出,突破双参量贝叶斯统计推断无法处理高维数据的局限性。经过实验验证,该系统能够有效提高集料级配检测效率及精度,针对合格集料的检测误差在±2.5%以内。展开更多
我国输电线路存在异常检测数据准确性和及时性较低,无线环境恶劣,数据在时空难关联等问题,因此建设一个高效、安全、准确的输电线路异常检测模型迫在眉睫。提出一种基于链形混合拓扑的异常检测方法,将传感器采集到的数据送至基站进行单...我国输电线路存在异常检测数据准确性和及时性较低,无线环境恶劣,数据在时空难关联等问题,因此建设一个高效、安全、准确的输电线路异常检测模型迫在眉睫。提出一种基于链形混合拓扑的异常检测方法,将传感器采集到的数据送至基站进行单源和多源多维数据异常检测。该方法首先设计了一种基于时间维度的单源数据异常检测算法(single-source data anomaly detection algorithm,SDADA),对检测时间内的数据进行依次遍历,确定有效和异常数据的个数,然后对异常检测结果进行综合分析。其次,设计了一种在基站端执行的多源多维数据异常检测算法(multi-source and multi-dimensional data anomaly detection algorithm,MDADA),在SDADA的基础上,通过位置相关性定义了不同传感器之间的距离关系,用于确定候选异常检测队列,并对特定时间的异常数据值进行综合分析。实验结果表明,与传统方案相比,该方法具有更高的检测精度和执行效率。展开更多
文摘Blood vessels in ophthalmoscope images play an important role in diagnosis of some serious pathology on retinal images. Hence, accurate extraction of vessels is becoming a main topic of this research area. In this paper, a new hybrid approach called the (Genetic algorithm and vertex chain code) for blood vessel detection. And this method uses geometrical parameters of retinal vascular tree for diagnosing of hypertension and identified retinal exudates automatically from color retinal images. The skeletons of the segmented trees are produced by thinning. Three types of landmarks in the skeleton must be detected: terminal points, bifurcation and crossing points, these points are labeled and stored as a chain code. Results of the proposed system can achieve a diagnostic accuracy with 96.0% sensitivity and 98.4% specificity for the identification of images containing any evidence of retinopathy.
文摘合理的集料级配能有效提升道路性能并延长道路使用寿命,针对拌合站现场需要快速进行集料级配检测的诉求,设计基于双参量MCMC算法的集料级配检测系统。针对单视角集料图像无法准确反映集料实际粒径的问题,该系统设置双相机进行图像采集;并引入基于凹点匹配的图像分割算法,解决副相机采集过程中出现的颗粒堆叠问题;为尽可能准确地表征集料颗粒等效粒径,经实验验证选择等效椭圆短径均值和等效Feret短径均值分别表征不同粒径区间内集料颗粒;将双相机采集图像中提取到的集料颗粒信息作为双参量输入,利用双参量贝叶斯统计思想对集料实际粒径分布进行推断;在工程上引入马尔科夫链蒙特卡洛(Markov chain Monte Carlo,MCMC)算法计算双参量贝叶斯后验分布,并将其作为集料级配结果输出,突破双参量贝叶斯统计推断无法处理高维数据的局限性。经过实验验证,该系统能够有效提高集料级配检测效率及精度,针对合格集料的检测误差在±2.5%以内。
文摘我国输电线路存在异常检测数据准确性和及时性较低,无线环境恶劣,数据在时空难关联等问题,因此建设一个高效、安全、准确的输电线路异常检测模型迫在眉睫。提出一种基于链形混合拓扑的异常检测方法,将传感器采集到的数据送至基站进行单源和多源多维数据异常检测。该方法首先设计了一种基于时间维度的单源数据异常检测算法(single-source data anomaly detection algorithm,SDADA),对检测时间内的数据进行依次遍历,确定有效和异常数据的个数,然后对异常检测结果进行综合分析。其次,设计了一种在基站端执行的多源多维数据异常检测算法(multi-source and multi-dimensional data anomaly detection algorithm,MDADA),在SDADA的基础上,通过位置相关性定义了不同传感器之间的距离关系,用于确定候选异常检测队列,并对特定时间的异常数据值进行综合分析。实验结果表明,与传统方案相比,该方法具有更高的检测精度和执行效率。
