Upload Code
loading-left
loading loading loading
loading-right

Loading

Profile
No self-introduction
codes (58)
A program that converts video into images
no vote
The purpose of the program is to convert the video into a frame of image process, using MATLAB, as long as the user modifies the path of the video file can run, the image results will be stored in the corresponding root directory.
fwm
2018-03-08
1
1
Infrared pedestrian test video
no vote
  Target detection in infrared video is much more difficult than that in visible light because of the quality of infrared video It is very easy to be affected by the equipment that takes infrared video. Once the performance of the equipment is not stable enough, the infrared video will bring great difficulty in the later detection process, so the infrared video shooting equipment has higher requirements than the visible light. This resource contains infrared video recorded by three infrared cameras, which can be used for infrared target tracking test.
fwm
2018-03-08
0
1
A program that converts an image into a video
no vote
The purpose of the program is to convert a frame of image into a video file. It is written in MATLAB. As long as the user modifies the storage path of the image file, it can run. The result of the video file will be stored in the corresponding root directory.
fwm
2018-03-08
1
1
Hog feature extraction
no vote
IMG: & nbsp; is the input image, cellpw and cellph are the pixel width and height of the cell, respectively. &Nblockw, nblcokh: the number of cells in the X and Y directions in the block. &Nbsp; & nbsp; nthet: is the number of gradient direction histogram containers & nbsp; issigned: is the condition that the gradient direction histogram ranges from 0 to PI without sign and 0 to 2 * PI with sign. It can be specified by setting the value issigned of the variable. &Nbsp; overlap: & nbsp; the ratio of two adjacent blocks to overlap. &Isglobalinterpolate: in strict accordance with the requirements of DALAL's paper. &Nbsp; normmethod: is the block histogram normalization method, set to the following string: 'none', which means denormalization; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; 'L1', which means L1 normalization; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; && nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; 'L2', which means L2 specification Standardization & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; & nbsp; &This means that the l1sqrt specification is standardized
fwm
2018-01-06
0
1
Gray histogram feature extraction with spatial location information
no vote
The traditional gray histogram lacks the spatial distribution information of pixels. If the spatial distribution of pixels can be included in the histogram, it will better describe the target. Considering that not all the pixels in the target region have the same contribution to the description of the target gray, but usually the closer the point is to the center of the target, the greater the contribution is. Therefore, we use a weighting function to give a smaller weight value to the pixels far away from the center of the target region.
fwm
2018-01-05
2
1
Using probabilistic data Association algorithm and
4.0
CV model, the probability data association algorithm and the nearest neighbor algorithm are used to track and filter them to ensure the accuracy; JPDA data association algorithm is used to realize the correlation between the point trace and the track of two uniform moving targets
fwm
2016-08-23
2
1
JPDA implements two underwater target data Associa
4.0
The JPDA data association algorithm is used to realize the correlation between the track and the track of the two uniform moving target; & nbsp; input variable & nbsp; & -target_position:; nbsp; the initial position of the target
fwm
2016-08-23
2
1
Meanshift combines color and texture to track targ
no vote
Application background Meanshift is first proposed by Comaniciu, is a kind of algorithm for tracking non rigid objects. In order to characterize the object, we need to select a feature space. The reference target is represented by the probability density function Q in the feature space, for example, the reference model can choose the color probability density function of the target. Without loss of generality, the target model can be considered as the center of the center of its space. In the next frame, the candidate target is defined as the position Y, which is represented by the probability density function p (y). Key Technology The algorithm combines color and texture information to track the target, which contains the video file to test. The similarity function between the target and the candidate histogram is denoted as a formula (1), and the local maximum of the similarity means that the candidate model is matched with the target model. If
fwm
2016-08-23
0
1
Particle filter based on frame difference method i
no vote
Application background The so-called particle filter is to find a set of random samples in the state space to approximate the probability density function, using the sample mean instead of integral calculation, and then obtain the system state of the minimum variance estimation process, the image of these samples is called "particle", so called particle filter. Key Technology Although the probability distribution of the algorithm is an approximate one, but because of the characteristics of the non parametric, it can get rid of the problem of nonlinear filtering. It must satisfy the Gauss distribution. It can be expressed more widely than Gauss model. Therefore, particle filter can be used to accurately express the posterior probability distribution of measurement and control, which can be used to solve the SLAM problem.
fwm
2016-08-23
1
1
Fusion color and shape feature particle filter for
no vote
Application background Particle filter technology in the non-linear, non - Gauss system performance of the superiority, it is determined that its application is very wide. In addition, the particle filter is one of the reasons for the application of multi mode processing capability. Internationally, particle filtering has been applied in various fields. In the field of economics, and it is applied in forecasting the economic data; in the military field has been applied in radar tracking flying object, empty to empty, air to ground passive tracking; in the field of traffic control it is applied in the car or video surveillance; it is also used in the robot global localization. Key Technology This program combines color histogram and histogram of the gradient of the video image processing, in the target and background to enter the case, because the addition of shape features can be a good tracking target, while the target is partially occluded in 4
fwm
2016-08-23
1
1
View More