排序方式: 共有6条查询结果,搜索用时 15 毫秒
1
1.
分数阶微积分在图像处理过程中,能够加强图像的纹理细节,也有利于图像的特征匹配。为了降低在分数阶图像匹配过程中寻找合适阶数所需要的大量人力和物力,本文提出了自适应的分数阶尺度不变特性变换(SIFT)算法,将自适应的图像处理方法与高斯滤波相结合,利用自适应的分数图像处理对图像加强,然后检测出更准确的关键点,进行图像匹配,提高图像的匹配率。实验表明自适应的分数阶在匹配中的应用减少了寻找合适阶数的麻烦,而且也可以达到较理想的匹配率。 相似文献
2.
针对目前的分数阶非线性模型图像特征提取能力不足导致分割精度较低的问题,提出一种基于分数阶网络和强化学习(RL)的图像实例分割模型,用来分割出图像中目标实例的高质量轮廓曲线。该模型共包含两层模块:1)第一层为二维分数阶非线性网络,主要采用混沌同步方法来获取图像中像素点的基础特征,并通过根据像素点间的相似性进行耦合连接的方式获取初步的图像分割结果;2)第二层通过RL思想将图像实例分割建立为一个马尔可夫决策过程(MDP),并利用建模过程中的动作-状态对、奖励函数和策略的设计来获取图像的区域结构和类别信息。最后将第一层获取到的像素特征和初步的图像分割结果与第二层获取到的区域结构和类别信息联合起来进行实例分割。在Pascal VOC2007 和Pascal VOC2012数据集上的实验结果表明,这种基于连续决策的图像实例分割模型与传统的分数阶模型相比,平均精度(AP)至少提升了15个百分点,不仅能够获取图像中目标物体的类别信息,而且进一步提升了对图像轮廓细节和细粒度信息的提取能力。 相似文献
3.
4.
5.
Multiphase nano-Ni(OH)2 doped with Y or La was prepared by supersonic co-precipitation method. The crystal morphology, structure and particle size were characterized by transmission electron microscopy (TEM), X-ray diffraction (XRD) and particle size distribution (PSD). The electrochemical performance of samples was investigated by electrochemical workstation and battery tested system. The results indicated that micro-morphology and grain size were changed with the changing of supersonic power, pH values and doping elements. The morphology of Y doped sample was from the flake-like to the needle-like with the increase of supersonic power; Particles were from quasi-spherical particles into needle-like with the increase of pH values; As the supersonic power increased, the proportion of α-Ni(OH)2 increased initially and then decreased. pH value was very important to the formation of crystalline phase. Lower pH value was beneficial to the formation of α-Ni(OH)2. However, the pH values had a slight effect on the reaction reversibility. Complex electrodes were prepared by mixing 8 wt.% nickel hydroxides with commercial micro-size spherical nickel. The discharge capacity of electrodes increased initially and then decreased with the increase of supersonic power. When the supersonic power was 60 W and the pH value was 9, the sample had the largest dis-charge capacity (358 mAh/g) at 0.5 C rate, which was 122.7 and 76 mAh/g higher than the spherical nickel electrode and La doped sample electrode, respectively. 相似文献
6.
超声波沉淀法制备Y掺杂纳米多相Ni(OH)2及其性能研究 总被引:1,自引:0,他引:1
采用超声波沉淀法制备了不同摩尔比例Y掺杂氢氧化镍,对其粒度分布、结构及电化学性能进行了测试分析。XRD测试表明,样品均为α和β相混合结构的Ni(OH)2。激光粒度测试表明,样品均为纳米颗粒,且分布均匀,平均粒径在50~80 nm之间。分别将制备的样品以8%比例与工业用微米级球镍混合制成复合镍电极,电极的可逆性和充电效率均随Y掺杂比例增大先提高后下降,Y含量 1.17%时,其电极可逆性和充电效率达到最佳,放电比容量达到最大,0.1和0.5 C倍率下的比容量分别达到370和358 mAh/g,且具有较低充电电压和较高放电平台,该结果比目前市售镍氢电池比容量(230~250 mAh/g)高48%~60%。对加超声波和不加超声波制备的样品性能进行了比较 相似文献
1