Saturday, September 13, 2008

How To Design A Boat To Hold Pennies

wavelet transformations of the sky background modeling

From the extraction of the bottom of a deep sky image we can correct the problems non-uniform illumination and color, such as those associated with vignetting. This forms part of own calibration of the image, and ideally should be applied on raw data linear.

PixInsight has two processing modules extraction mechanisms and automatic dynamic background:
  • AutomaticBackgroundExtraction (ABE)
  • DynamicBackgroundExtraction (DBE)

In normal situations ABE get an excellent modeling of the background of the image, while DBE is more flexible and appropriate when faced with more complicated cases, especially when there is much to refine over the procedure, and the user can intervene dynamically over the image.

Depending on the nature of the problem in different lighting and / or color, image correction must be made by subtraction or division. The vignetting for example is a multiplier effect, so the division is more appropriate. On the other hand, light pollution and atmospheric extinction are additive effects, so that should be neutralized by subtracting.

This is a case of vignetting easy to correct with ABE by division. I have used most of the parameters included by default, except under the generation and global decline of samples, thus avoiding some of the boxes are included on the photographed object.

Figure 1 .- Original vignetting

Figure 2 .- Distribution of samples in ABE

Figure 3 .- Extraction Automatic background (pseudo-flat)

Figure 4 .- Correction to ABE by division

DBE is instead appropriate in more complex cases, as this wide-field photography affected by atmospheric extinction and light pollution. In this test uses only the default, which already introduce a significant improvement.


Being able to act dynamically on the image allows the user to manually place the samples on the bottom of the sky and forcing the modeling mechanism in difficult situations, such as gradients themselves Milky Way.

Friday, September 12, 2008

Can You Take Lorazepam And Benadryl

RGB Color Calibration

The color photographs are almost always required if we are to be calibrated a good color balance across the image. Due to the characteristics of the capture devices, either digital or CCD , the answer is usually not uniform throughout the color spectrum. It is therefore necessary to adjust the weight of each RGB channel . And of course this setting must be linear, because otherwise we can not get a perfect color balance throughout the image.

I always performed the color calibration in PixInsight with the processing module ChannelMatch . My argument is based primarily on the statistical values \u200b\u200bof an object which I know their color. This object reference is usually the sky background, which is assumed to be completely neutral, with no color casts.


First visualize raw images helping a linear transfer function display ( ScreenTransferFunction ) and thus easily locate the regions of sky. Then generated from the initial data from a separate image a preview , which usually call ' background. " Reading and histogram representation of the background shows me clearly indicates that the distribution of pixel values \u200b\u200bof each color channel differ greatly.


The idea is to overlap the bells for each channel of the histogram from ChannelMatch modifying linear correction factors. This way I get a correct color balance, since the main peak of the distribution is often a good indicator of one's own sky.


As you can see above, by varying the correction factors, the bells of the histogram of each color channel are perfectly overlapped, thus obtaining a sample of completely neutral background.


finally obtained correction factors (in this case R = 0.85, G = 0.66 and B = 1.0) are applied over the entire image.

Thursday, September 11, 2008

Gay Cruising Spots In Northern New Jersey

nonlinear transformation (MTF)

When we get a series of photos from your digital camera or CCD first need to perform a calibration. The calibration process requires proper application darks, bias flats and light sockets on the photographed object, keeping the linear response of the capture device of the camera.

Through a nonlinear transformation of the data after the calibration process can be adapted to make the image more visible to our eyes, as the data is originally represented a small portion of the available dynamic range in the shadows, so that through a device like a monitor can only visually distinguish saturated image pixels.

This nonlinear transformation is basically a transfer function halftone, which lets you change the values \u200b\u200bdepending on brightness of each pixel.

The following figure shows the histogram of input (raw linear data) and output (linear distribution) HistogramTransformation generated from a deep-sky image.

linear distribution

Distribution nonlinear

From non-linear distribution, the information is distributed over the entire dynamic range available , allowing display devices show much more information recorded clearly.

Wednesday, September 10, 2008

I Wanna See My Cousins Bpobs



I created this blog mainly because I like to process astronomical images. My intention is to introduce some key processing techniques and update the blog with new techniques as they are developed. All this including some practical examples with the results of applying the various processes on the images.

My favorite application for image processing is PixInsight , a modular processing environment specializing in astrophotography. With this software I can have absolute control of every aspect in the treatment of a recorded image through any capture device, while allowing me to also develop creative skills.

I hope this blog will be useful for all the fans who, like astronomy, I enjoy getting your own photographs, both in solar system objects deep sky.