matlab code

应用背景 SIFT特征匹配算法, 在没有先验知识条件下, 对旋转、尺度缩放、光照、视角变化等图像变化因素保持一定鲁棒性, 且对低重叠区域(30%左右的相似重叠区域)、部分遮挡(只出现在其中一张图上的物体)和噪声等因素也保持较好的可匹配性。可用于多尺度下的特征匹配，以及鲁棒性的图像拼接。 关键技术 1.高斯尺度空间极值检测 2.特征点位置的确定 3.特征点方向的分配 4.生成SIFT特征向量 5.图像拼接 6.图像融合 提取图像的sift特征，构造特征描述子，基于sift特征构造BBF-tree，根据描述子之间的欧式距离进行初步的匹配，然后进过鲁棒的Ransac估计出最佳运动模型H，然后用距离加权的方法在重叠区进行融合。

掌上电脑上的画图系统程序.-handheld computer systems on the drawing process.

数学曲面的绘制程序-a protract procedure foe mathmatic band face

本文介绍一种简单的装置，能使其实现对太阳水平、垂直方向的全方位跟踪，晚上便自动复位-solar tracking device

C++读写BMPJPG图像的方法,非常清晰的程序解释,清楚,可转移

直线扫描算法rivate AX1(100) As Integer, AY1(100) As Integer, AX2(100) As Integer, AY2(100) As IntegerPrivate CX1 As Integer, CY1 As Integer, CX2 As Integer, CY2 As Integer, N As IntegerPrivate Sub form_Load()Picture1.BackColor = RGB(255, 255, 255)N = 0End SubPrivate Sub picture1_MouseDown(Button As Integer, Shift As Integer, X As Single, Y As Single)If Button = 1 ThenCX1 = CInt(X): CX2 = CInt(X)CY1 = CInt(Y): CY2 = CInt(Y)End IfEnd Sub-linear scanning algorithm rivate AM1 (100) As Integer, AY1 (100) As Integer, AX2 (100) As Integer, AY2 (100) As IntegerPrivate CX1 As Integer, CY1 As Integer, CX2 As Integer, CY2 As Integer, N As IntegerPrivate Sub form_Load () Picture1.BackColor = RGB (255, 255, 255) N = 0End SubPrivate Sub picture1_MouseDown (Button As Integer, Shift As Integer, As Single X, Y As Single) If Button = = a ThenCX1 CInt (X) : CX2 = CInt (X) = CY1 CInt (Y) : CY2 = CInt (Y) End Sub IfEnd

用vc实现的联机手写数字识别程序 可以运行 很好的-By vc realize online handwritten numeral recognition process can run very good

人脸定位实例全部源码，包括对眼睛，嘴巴等的定位-face location of all source code examples, including the eyes, mouths positioning

freeimage library

园的扫描转换算法-Park scan conversion algorithm