Digital image processing using Histograms:


Histogram-Based Enhancements


  • The histogram is a graphical representation that brings a visual impression of the distribution of data. Rather
  •  than looking just at digital numbers, which becomes difficult to understand how the pixel values are 
  • distributed against an image. But if we plot a histogram,
    which is easy to do, on the y-axis we get the frequency on the x-axis we get the pixel value. And once it 
  • is done, then a lot of information as also discussed in previous lectures can be gathered about a particular
  •  image. So histograms provide a basically univariate because histograms of individual bands, individual 
  • images there, that is why it is univariate. And if we want to see the histogram of 2, then it would be bivariate
  •  or multivariate or also possible. So one variable description of data, how data is distributed and that tells
  •  also how the quality of an image taken by a sensor, so that can be seen An image histogram is basically 
  • which is a not only graphical representation, but also it tells you that whether there are lighter objects or
  •  darker objects present in the image and basically it plots each pixel have that image Histogram, why 
  • histogram needs to be done or studied, assessed, that is the basic information it provides. And of which is
  •  a start of image processing or image enhancement techniques. And because there are various spatial 
  • domain processing steps we might be taking. And to understand or choose the best one, first, we have to 
  • understand the histogram.sometimes should always be spent on image








  • histogram and statistics, before we go for any kind of image enhancement techniques. When we enhance 
  • images basically what we are doing we are manipulating with the histogram and we will be seeing some
  •  examples also. And histogram provides the information about the image statistics as well mean, median,
  •  and mode. Minimum value, maximum value what is the maximum frequency that kind of information can 
  • be retrieved with the histogram. If there are 2 peaks, it tells something else, if there is a single peak
    that also, if the distribution is Gaussian distribution, then it tells a different thing, so that is why it is important
  • . And the information derived from histograms is quite useful in many other applications of digital image 
  • processing or in remote sensing may be in image compression and segmentation. So, image compression,
  •  why because this image compression is required to reduce the size of an image without compromising the 
  • quality of the image, is the purpose here for us. And if there is a homogeneity present in the image, a lot
  •  of compression image compression can be achieved. And if there is a lot of heterogeneity present among 
  • the pixel values then high image competition cannot be achieved. 

  • So, if I study a histogram, I can assess whether the distribution is heterogeneous or homogeneous. If it is 
  • homogeneous, I know now, that image will provide a particular compression technique that will provide me 
  • better results. So, that is why it can be used also for image segmentation in the same way it can be used.
  •  Though understanding histograms have got various applications. Now, basically, it is a frequency distribution
  •  in a bar graph that displays how often observed
  • values fall within certain intervals or classes. So, these towers which you see are basically intervals or
  •  different classes are there. We can also control the number of classes, and how this thing should be
  •  displayed. And relative portions of data that fall in each class are represented by the height of each building
  •  or bar. So, for example, the histogram is shown here, and that is showing basically frequency distribution in
  •  10 classes. So, it depends on my requirements I want more number of classes or even if it is the 8-bit 
  • scenario. Then I can have even 256 classes, with variations between 0 to 255. And I can see the things 
  • accordingly, same time also you get the minimum value, maximum value, mean, mode, median, standard
  •  deviation. Another statistical parameters like skewness, kurtosis, and all those things can also be identified
  •  from there, so that is why histogram is important. Histogram brings measures of a step basically how pixel 
  • values are distributed in an image, spread of points around the mean is another characteristic of display 
  • histogram or frequency distribution. As you can see that the red curve is showing less distribution around
  •  the mean whereas
  •  the white curve is showing more distribution around the mean. And the variance and standard deviation for
  •  the black frequency distribution as you can see are greater than those of the red frequency distribution.
  •  If you are having skewness present in your image, some a bright part or more or dark part is more. Then 
  • in the histogram, you would see a skewed representation skew distribution in the image,




  • so that can be also the shape of the histogram that tells about the skewness present. And this coefficient of
  •  skewness is basically a measure of the symmetry of distribution. So, generally, you do not get a symmetric 
  • histogram generally you get a systematic histogram of distribution. And therefore, there will be skewness 
  • either towards you know smaller values or large values. In this example skewness towards the lower
  •  value and therefore, mean, median, and mode will be located at different places. So, that also tells that
  •  there are alike an I can while going through this histogram, I can tell that. There is a large number of pixels 
  • which are having lower pixel values compared to the values which are having high pixel values. And if it
  •  would have been reversed, then a different interpretation will come out from this one.

Comments

Post a Comment

Popular posts from this blog

Curriculum Vitae

Image
Muhammad Shahid Usman Mobile:+92 346 4500427  Email:shahidusmaninfo@gmail.com    OBJECTIVES: A Suitable position with an organization where I can utilize the best of my skills and abilities that fit my education, skills, and experience a place where encourage and permitted to be an active participant as a well vital contribution to the development of the company. PERSONAL DETAILS: ·          Name:                               Muhammad Shahid Usman ·          Father Name:               Muhammad Munir Khokhar ·          Date of Birth:                16/03/2002 ·          Identity Card:         ...
Read more

Atmospheric Errors in satellite Image

Image
  Effects of Atmosphere on digital Image processing: Hello everyone, and welcome to our new discussion which is on atmospheric errors and corrections under this remote sensing essential course. Earlier we briefly touched on this issue and but today we are going to discuss in depth the reasons for atmospheric distortions and then how these are reflected in our satellite images finally we will also see what are the  possibilities of getting rid of atmospheric distortions.  So as you know that you know between satellites and earth surface there is atmosphere. But if the atmosphere is very thin or would have been thin then there should have been a delight for RCO or for any need remote sensing people. But unfortunately, it is not true in the case of earth and though it is very important for life, it is fine. But if we think about mars, then mars is having a very thin atmosphere. And therefore though it is written here the geologist delight but does not mean only for geologist...
Read more