-
kalman-filter
说明: 经典卡尔曼滤波器代码。对初学者十分有帮助。(Classic kalman filter code, for beginners introduction is very helpful.)
- 2011-02-20 11:25:41下载
- 积分:1
-
sndr
计算求解连续时间信号的SINAD,分析其频谱特型(Calculate the SINADof Real Signal)
- 2013-07-27 16:25:47下载
- 积分:1
-
aliasing
alaisin concept in communication system
- 2013-03-06 17:18:56下载
- 积分:1
-
irisonedemo
work in matlab, iris detection hamming distance
- 2013-10-22 09:16:31下载
- 积分:1
-
kalmanfilter
Matlab code for Kalman filter
- 2013-09-27 02:23:14下载
- 积分:1
-
cost207
MIMO Channel model cost207 file.
- 2015-02-18 17:43:48下载
- 积分:1
-
BPNN
反向传播神经网络的学习算法以及MATLAB实现代码(Back propagation neural network learning algorithm and MATLAB code)
- 2015-03-19 20:28:42下载
- 积分:1
-
daolibai_Fuzzy_Controller
基于模糊算法的倒立摆控制的matlab仿真源程序(Inverted pendulum fuzzy control algorithm matlab simulation source based)
- 2017-01-14 09:34:26下载
- 积分:1
-
videoface
说明: 直接使用摄像头检测人脸的程序,正确率还不是很准确(Directly face the camera detection procedure, the correct rate is not very accurate)
- 2011-02-19 05:59:29下载
- 积分:1
-
sift
1 SIFT 发展历程
SIFT算法由D.G.Lowe 1999年提出,2004年完善总结。后来Y.Ke将其描述子部分用PCA代替直方图的方式,对其进行改进。
2 SIFT 主要思想
SIFT算法是一种提取局部特征的算法,在尺度空间寻找极值点,提取位置,尺度,旋转不变量。
3 SIFT算法的主要特点:
a) SIFT特征是图像的局部特征,其对旋转、尺度缩放、亮度变化保持不变性,对视角变化、仿射变换、噪声也保持一定程度的稳定性。
b) 独特性(Distinctiveness)好,信息量丰富,适用于在海量特征数据库中进行快速、准确的匹配[23]。
c) 多量性,即使少数的几个物体也可以产生大量SIFT特征向量。
d) 高速性,经优化的SIFT匹配算法甚至可以达到实时的要求。
e) 可扩展性,可以很方便的与其他形式的特征向量进行联合。
4 SIFT算法步骤:
1) 检测尺度空间极值点
2) 精确定位极值点
3) 为每个关键点指定方向参数
4) 关键点描述子的生成
本包内容为sift算法matlab源码(1 SIFT course of development
SIFT algorithm by DGLowe in 1999, the perfect summary of 2004. Later Y.Ke its description of the sub-part of the histogram with PCA instead of its improvement.
2 the SIFT main idea
The SIFT algorithm is an algorithm to extract local features in scale space to find the extreme point of the extraction location, scale, rotation invariant.
3 the main features of the SIFT algorithm:
a) SIFT feature is the local characteristics of the image, zoom, rotate, scale, brightness change to maintain invariance, the perspective changes, affine transformation, the noise also maintain a certain degree of stability.
b) unique (Distinctiveness), informative, and mass characteristics database for fast, accurate matching [23].
c) large amounts, even if a handful of objects can also produce a large number of SIFT feature vectors.
d) high-speed and optimized SIFT matching algorithm can even achieve real-time requirements.
e) The scalability can be very convenient fe)
- 2012-05-25 15:31:16下载
- 积分:1