Description: 一种图像拼接源码。该方法首先运用基于列特征的匹配方法,确定出图像重叠区域;然后采用直方图匹配方法进行灰度调整;最终使用加权平滑算法完成图像的无缝拼接。-A source image mosaic. First of all, the method characterized by the use of column-based matching method to determine the image overlap region then histogram matching methods used greyscale adjustment end-use-weighted smoothing algorithm to complete a seamless image mosaic. Platform: |
Size: 1024 |
Author:李永冰 |
Hits:
Description: 自己编的一些关于图像灰度直方图处理的MATLAB源程序,包括加权距离,累加直方图,欧氏距离,直方图相交法,中心距法共5种方法,可选择适合的选用。-Own some on the image histogram processing MATLAB source code, including the weighted distance, cumulative histogram, Euclidean distance, histogram intersection of law, from the Law Center, five kinds of methods to choose the appropriate selection. Platform: |
Size: 2048 |
Author:louwutao |
Hits:
Description: GUI界面,能方便的对图像进行灰度变换、直方图均衡、增大及缩小分辨率、不同模板数的高斯平滑滤波器和加权平滑滤波器等
-GUI interface, can make it easier for the gray-scale transformation of the image, histogram equalization, increase and reduce the resolution, the number of different templates Gaussian smoothing filter and the weighted smoothing filters, etc. Platform: |
Size: 804864 |
Author:王勇 |
Hits:
Description: 针对目前的基于特征的图像检索中没有有效地结合图像中对象空间信息的问题,提
出了一种新的融合了颜色、空间和纹理特征的图像特征提取及匹配方法。为了减少时间
间复杂度,首先通过基于普通颜色直方图的检索得到初始图像集合,然后根据提出的结合空间、纹理特征加权度量对初始图像集合再进行检索,从而得到最后更符合要求的相似图象-View of the current feature-based image retrieval is not effective integration of image information of objects in space, we propose a new blend of colors, space and texture features of image feature extraction and matching method. Time interval in order to reduce complexity, first of all, the general color histogram-based retrieval has been the initial image set, and then on the basis of the combination of space, texture characteristics of the weighted measure of the initial image and then search the collection, thus more in line with the requirements of the final similarity Image Platform: |
Size: 9111552 |
Author:丁丁 |
Hits:
Description: 加权系数k均值算法,自动确定初始类中心,基于图像直方图的快速k均值算法-Weighted coefficient of k-means algorithm automatically determines the initial type of center, based on image histogram fast k-means algorithm Platform: |
Size: 1024 |
Author:刘炎 |
Hits:
Description: 图像检索中颜色的特征提取及匹配算法,以家权欧几里得距离,中心距得加权距离,直方图交集算法等。-Image Retrieval color feature extraction and matching algorithm to the right home Euclid distance, center distance of a weighted distance, histogram intersection algorithm. Platform: |
Size: 35840 |
Author:李里 |
Hits:
Description: 本文提出了一种新的基于对模型的分块直方图求交互信息量的镜头检测算法。算法中对每帧图象
进行分块求直方图, 然后利用相邻帧间对应分块的直方图统计值求交互信息量, 最后把所有分块的交互信息量进行
加权平均以检测镜头的变化。实验结果表明与传统的直方图相比, 该算法对一般场景的镜头变化有更高的查全率和
查准率。-This paper presents a new model based on the sub-block histogram information for cross-detection algorithm of the camera. Algorithm for each image frame block for histogram, and then frame the use of adjacent sub-block histogram corresponding statistics for the amount of information interaction, and finally all of the interactive block to the weighted average amount of information to detect changes in lens . The experimental results show that compared with the traditional histogram, the algorithm for general scenes of the lens changes in higher recall rate and precision. Platform: |
Size: 349184 |
Author:丁金金 |
Hits:
Description: fit_maxwell_pdf - Non Linear Least Squares fit of the maxwellian distribution.
given the samples of the histogram of the samples, finds the
distribution parameter that fits the histogram samples.
fits data to the probability of the form:
p(r) = sqrt(2/pi)*(a^(-3/2))*(r^2)*exp(-(r^2)/(2*a))
with parameter: a
format: result = fit_maxwell_pdf( x,y,W,hAx )
input: y - vector, samples of the histogram to be fitted
x - vector, position of the samples of the histogram (i.e. y = f(x,a))
W - matrix or scalar, a square weighting matrix of the size NxN where
N = length(y), or 0 to indicate no weighting is needed.
hAx - handle of an axis, on which the fitted distribution is plotted
if h is given empty, a figure is created.
output: result - structure with the fields
a - fitted parameter
VAR - variance of the estimation
type- weighted LS or not weighted LS
iter- number of iteration for the solution- fit_maxwell_pdf - Non Linear Least Squares fit of the maxwellian distribution.
given the samples of the histogram of the samples, finds the
distribution parameter that fits the histogram samples.
fits data to the probability of the form:
p(r) = sqrt(2/pi)*(a^(-3/2))*(r^2)*exp(-(r^2)/(2*a))
with parameter: a
format: result = fit_maxwell_pdf( x,y,W,hAx )
input: y - vector, samples of the histogram to be fitted
x - vector, position of the samples of the histogram (i.e. y = f(x,a))
W - matrix or scalar, a square weighting matrix of the size NxN where
N = length(y), or 0 to indicate no weighting is needed.
hAx - handle of an axis, on which the fitted distribution is plotted
if h is given empty, a figure is created.
output: result - structure with the fields
a - fitted parameter
VAR - variance of the estimation
type- weighted LS or not weighted LS
iter- number of iteration for the solution Platform: |
Size: 2048 |
Author:resident e |
Hits:
Description: fit_rayleigh_pdf - Non Linear Least Squares fit of the Rayleigh distribution.
given the samples of the histogram of the samples, finds the distribution parameter that fits the histogram samples.fits data to the probability of the form: p(r)=r*exp(-r^2/(2*s))/s
with parameter: s
format:result = fit_rayleigh_pdf( x,y,W,hAx )
input: y - vector, samples of the histogram to be fitted
x - vector, position of the samples of the histogram (i.e. y = f(x,a))
W - matrix or scalar, a square weighting matrix of the size NxN where
N = length(y), or 0 to indicate no weighting is needed.
hAx - handle of an axis, on which the fitted distribution is plotted
output: result - structure with the fields
s - fitted parameter
VAR - variance of the estimation
type- weighted LS or not weighted LS
iter- number of iteration for the solution-fit_rayleigh_pdf - Non Linear Least Squares fit of the Rayleigh distribution.
given the samples of the histogram of the samples, finds the distribution parameter that fits the histogram samples.fits data to the probability of the form: p(r)=r*exp(-r^2/(2*s))/s
with parameter: s
format:result = fit_rayleigh_pdf( x,y,W,hAx )
input: y - vector, samples of the histogram to be fitted
x - vector, position of the samples of the histogram (i.e. y = f(x,a))
W - matrix or scalar, a square weighting matrix of the size NxN where
N = length(y), or 0 to indicate no weighting is needed.
hAx - handle of an axis, on which the fitted distribution is plotted
output: result - structure with the fields
s - fitted parameter
VAR - variance of the estimation
type- weighted LS or not weighted LS
iter- number of iteration for the solution Platform: |
Size: 2048 |
Author:resident e |
Hits:
Description: 量化的hsv颜色直方图 不等间隔分为72维
效果还可以 注释里有加权直方图 但是效果不好-Quantify the hsv color histograms ranging interval is divided into 72-dimensional effects can also be annotated with the weighted histogram but ineffective Platform: |
Size: 1024 |
Author:张惟中 |
Hits:
Description: 背景建模和前景分割的方式把运动车辆提取出来。并进行最近临关联,输出目标轨迹。
MeanShift运动目标跟踪 matlab程序
1.截取跟踪目标矩阵rect
2.求取跟踪目标的加权直方图hist1
3.读取视频序列中的一帧, 先随机取一块与rect等大的矩形,计算加权直方图hist2。
4.计算两者比重函数,如果后者差距过大, 更新新的矩阵中心Y,进行迭代(MeanShift是一种变步长可以迅速接近概率密度峰值的方法),直至一定条件后停止。
这两种算法为处理图像跟踪的基本方法
-Background modeling and foreground segmentation the extracted moving vehicles. Pro associated output and the target locus. Target tracking matlab program MeanShift movement. Interception tracking the target matrix rect strike a tracking target weighted histogram hist1 3. Reads a video sequence, randomly take a with the rect large rectangular, calculate the weighted histogram hist2. 4. Calculate both the proportion of function, if the latter gap is too large to update the new center of the matrix Y conducted the iteration (MeanShift is a variable step size can be rapidly approaching the peak of the probability density) until certain conditions to stop. The basic method of these two algorithms for processing the image tracking Platform: |
Size: 240640 |
Author:l刘翔 |
Hits:
Description: 背景加权直方图算法(BWH)在[2]中提出了尝试
减少干扰的背景均值漂移跟踪的目标定位。然而,
在本文中,我们证明了权重分配给候选目标区域的像素
BWH是那些没有背景资料成正比,即不会引入BWH
任何新的信息,因为均值漂移迭代公式是不变的规模
改造砝码。然后,我们提出了一个校正BWH(CBWH)的公式
只转型的目标模式,但不是目标候选模型。 CBWH计划
可以有效地降低背景的干扰,在目标定位。实验
结果表明,CBWH可能会导致更快的收敛速度和更准确的定位比
通常的目标表示均值漂移跟踪。即使目标没有得到很好的初始化,
该算法仍然强劲跟踪的对象,这是很难实现由
传统的目标表示。-The background-weighted histogram (BWH) algorithm proposed in [2] attempts to
reduce the interference of background in target localization in mean shift tracking. However,
in this paper we prove that the weights assigned to pixels in the target candidate region by
BWH are proportional to those without background information, i.e. BWH does not introduce
any new information because the mean shift iteration formula is invariant to the scale
transformation of weights. We then propose a corrected BWH (CBWH) formula by
transforming only the target model but not the target candidate model. The CBWH scheme
can effectively reduce background’s interference in target localization. The experimental
results show that CBWH can lead to faster convergence and more accurate localization than
the usual target representation in mean shift tracking. Even if the target is not well initialized,
the proposed algorithm can still robustly track the object, which is hard to achieve by the
conventiona Platform: |
Size: 730112 |
Author:吴盈 |
Hits:
Description: MeanShift运动目标跟踪 matlab程序
思路:
1.截取跟踪目标矩阵rect
2.求取跟踪目标的加权直方图hist1
3.读取视频序列中的一帧, 先随机取一块与rect等大的矩形,计算加权直方图hist2。
4.计算两者比重函数,如果后者差距过大, 更新新的矩阵中心Y,进行迭代(MeanShift是一种变步长可以迅速接近概率密度峰值的方法),直至一定条件后停止。
-MeanShift moving target tracking matlab program ideas: 1. Interception target tracking matrix rect 2. Strike target tracking weighted histogram hist1 3. Reading of a video sequence, first randomly selected one with such a large rectangle rect, calculate the weighted Histogram hist2. 4 Calculate the proportion of both functions, if the latter gap is too large, updated new matrix center Y, iterative (MeanShift is a variable step size can rapidly approaching the peak of the probability density method), until certain conditions stop. Platform: |
Size: 2048 |
Author:LGF |
Hits: