A complete guide to astronomical image processing (3): Basic technology of deep space celestial bodies

Part III: Deep Space Objects-Basic Technology

In this section, you will learn:

> removing light pollution

& gt histogram

& gt printing color

& gt enhancement technology

The third part will jump out of the solar system and begin to introduce the processing methods of deep space objects. The software and tools used are also different from the first two software and tools dealing with the moon, planets and the sun.

Photos of deep space objects are often obtained through long-term exposure, which is not so simple to process. Light pollution often forms orange-red background pollution on photos. This part will talk about how to remove the background caused by this light pollution and how to extract every detail from the image.

This article will also introduce some skills that can make your color image clearer, sharper and shorten the shooting time, and how to view an image with histogram conveniently. Understanding histogram is the key to find hidden details in images.

We will also discuss the sharpening of images, how to sharpen the parts that need sharpening and keep the other parts unchanged. Finally, learn how to use layer mask to synthesize photos with different exposures.

Important software

Photoshop:Photoshop has always been the best graphic editing software on the market, but it is expensive. If you can't afford it, you can also use the simplified version of Photoshop Elements, which has enough functions. For the citizens of China, you can ignore this statement.

GIMP: the abbreviation of "GNU image control program" and a powerful editor based on layers. Different from the above two modes, it is free, supports various operating systems and has a large number of plug-in libraries.

Astronomical image processing software: Many astronomical image processing software are very easy to use, such as the powerful Maxim DL, Astroart and ImagesPlus, as well as the free IRIS.

Methods/steps

Remove light pollution

The starry sky is the most beautiful only when it is completely dark, and this condition is getting farther and farther away from us. Street lamps, floor lamps, car lamps and searchlights are all sources of light pollution, which destroys the starry sky in the city. The starry sky exposed for a long time will show a red hue, which is the result of light reflected by dust in the atmosphere.

There are several ways to avoid light pollution: shooting with monochrome CCD and narrow-band filter, or SLR and optical loss filter. If there is no filter, you can also improve the image during image processing.

Step 1: This picture is a typical feature caused by light pollution-the gradual change of orange-brown brightness.

Step 2: If the image has no gradient, it will be simpler. Open the color code, select the red channel, and slide the middle adjustment slider to the right until it appears a little green.

Step 3: Select the green channel again and slide the middle slider to the right until the sky background turns black again and there is no obvious red and green.

Step 4: If your image has obvious gradient, copy the layer, and then the two layers have the same image.

Step 5: In the above layer, select the filter/noise/dust and scratches, set the radius value to 100 and the threshold value to 0. This is, I can't see all the details.

Step 6: Change the blending mode of this layer to "poor", adjust the curve appropriately, and finally merge the layers.

bar chart

In the process of processing images, histograms are always encountered. Histogram is very useful in image processing, which describes the number of pixels of each brightness value in the image. The horizontal axis is brightness, the left is black, and the right is white. All brightness values in the image are arranged in sequence on the horizontal axis. The vertical axis is the number of pixels, because there are many brightness values, so the pixel values of various brightness are closely arranged together, showing some trends, like undulating mountains. The absolute number of pixels of a certain brightness is not important.

Histogram also provides a lot of content information of the image. For example, if an image has many stars on a black background, the histogram will have a peak on the far left, which is the information of the sky background.

The number of gray levels of the image taken by the camera is given by the number of bits. The more bits, the more gray levels. Therefore, many CCD cameras have high digits, and the gray scale series can reach thousands. This situation is not uncommon. For this kind of image, when it is opened in ordinary processing software, the histogram is almost all in the black end. Let's use histogram expansion technology to display images hidden in black.

Brightness and contrast expansion

Increasing brightness will shift the histogram of the image to the right. When moving too much, the bright pixel will reach the maximum, which is pure white. If you click OK at this time, these white pixels will not be restored.

Similarly, if the brightness is reduced until the histogram moves to the left, some pixels will become pure black, and the pure black background may look unnatural in deep space celestial photos. So this situation should be avoided.

The brightness distribution in the image, that is, the "main peak" in the histogram, defines the overall contrast of the image. For images with poor contrast, the "main peak" of the histogram is very narrow. The contrast is adjusted to widen this "peak".

Color gradient

The color scale tool adjusts directly on the histogram, and there are three sliders to adjust. Careful observation of the histogram can determine where adjustment is needed. Ideally, the black slider should be just before the histogram starts to rise, and the white slider should be after the histogram completely falls. At this time, its contrast is the best.

The middle slider defines the median brightness of the image. By adjusting this slider, you can change the brightness of the image without any pixels reaching the extreme value.

The color scale tool is very useful for extracting dark details, but it also has limitations, that is, linear adjustment. This means that you can't just change certain brightness and keep other brightness unchanged.

curve

The curve tool is a more powerful color scale. The color code tool has only one middle slider, while the curve theoretically provides an infinite number of middle sliders, which is a nonlinear adjustment. This can only be adjusted for certain brightness.

It is very simple to use. The curve starts from a straight line of 45 degrees. Generally, the horizontal axis represents the brightness to be adjusted, and the vertical axis represents the height to which the original brightness value is adjusted, that is, the outline of the curve. The default 45-degree line means that the brightness to be adjusted is the same as the target to be adjusted. You can redefine the shape of a curve by adding adjustment points to it and then dragging them.

The curve should be smooth to avoid the curve reaching the top or bottom, which means that some pixels become pure white and pure black.

Digital printing

Extracting details from an image is like developing a photo from a film. This is the so-called digital printing process (DDP). An important part of DDP is nonlinear histogram expansion.

DDP can be found in professional astronomical image processing programs, and one part of the histogram can be manually adjusted while keeping the other parts unchanged. This is useful in many cases, for example, if you want to enhance the faint and subtle structure in a galaxy or nebula without making other bright parts brighter.

Different software has different DDP implementation methods. In Maxim DL (below), DDP is a combination of USM and gamma extensions. The latter can enhance the bright or dark parts of the image.

Dealing with color-a time-saving method of monochrome photography

Many cameras can be used to shoot deep-space celestial bodies. Both SLR cameras and refrigeration CCD can shoot color photos directly, but we prefer to use monochromatic astronomical CCD. Although different filters are needed to shoot color photos, the shooting time is greatly prolonged.

The advantage of this is that you can separate the color information from the spatial details of the image, so that you can enhance the details of a certain color at will. The spatial details of an image are determined by brightness (L) information, which can be captured by L filter and color information can be captured by RGB filters. These three-color images are not as important as L images, and they do not improve the quality of image details. It sounds strange, doesn't it? That's true. If you have a very good L-image, it doesn't matter that the tricolor image can even be blurred.

When using CCD to shoot L images, it is necessary to set the resolution to the highest level to ensure that the best details can be shot. When shooting RGB three-color images, in order to achieve the exposure equivalent to L image, the exposure time should be three times as long as L image ideally. Because each color filter only transmits one third of the light.

However, because the resolution of three-color images is not important, the Ning Bin function of CCD can be used to improve the sensitivity of the camera, thus shortening the shooting time. Setting the Ning Bin with 3×3 pixels will make the sum of every 3×3 pixel become a large pixel, which will increase the sensitivity of the camera by 9 times and reduce the resolution by 1/9. So after using Ning Bin, the shooting time of each color image becomes one third of that of L image, and the result is the same as that of your original exposure of L image three times.

Steps:

The first step, suppose you have processed four images of L, R, G and B. If you take a three-color image by Ning Bin method, you need to increase the size of the image, just like the L image. If you use 2×2 Ning Bin, enlarge the image to 200%, and if you use 3×3 Ning Bin, enlarge it to 300%.

Step 2: Create an RGB image with the same size as the L image, paste four LRGB images into different layers, and name the layers to prevent mistakes. Hide the upper two layers and align the lower two layers.

Step 3: Repeat the previous step and align the upper two layers.

The fourth step is to create a new RGB image and paste the R layer into the red channel of the new image. Also paste the G layer and the B layer into the corresponding channels. Copy this full-color image (layer) and paste it back to the original image.

Step 5, move the L layer to the RGB layer, and set the layer blending mode to "lightness". Adjust saturation, color scale and curve until you are satisfied. If the RGB layer has noise that affects vision, you can do weak Gaussian blur.

In the sixth step, the L layer can also adjust the contrast, brightness, color scale and curve, and can also sharpen and enhance the details. The most important thing is to ensure the high quality of this layer. So it is necessary to copy this layer and hide it for backup.

Color adjustment technology

Add Hα information

To add the image shot by the Hα filter to the RGB image to enhance the details of Hα, on the basis of the above six steps, copy the R layer on the top layer, set the opacity to 30-50%, and then replace this layer and L layer with the Hα image.

Red enhancement

Open an RGB image, copy and paste the red channel into a new layer. Open the menu Image-Adjust-Hue/Saturation, check "Color", set the hue to 0 and the saturation to 50-70, then set the layer blending mode to "Brighten" and adjust the opacity until you are satisfied.

Substitute color

The substitute color tool can solve the purple edge problem. Turn on the Replace Color option, use the color picker to select the color of the purple edge, and select "Selection" below the thumbnail to increase the tolerance until the shape of the purple edge is displayed. Select black as the replacement color, and then fine-tune the saturation and brightness until the purple edge disappears.

Adjust the overall color balance

The color balance tool tool can adjust the overall color of the image. Color balance adjusts the color levels of three areas-shadow, midtone and highlight. For example, if the background of the image is too dark, select Shadow, and then adjust the sliders of three colors.

Deep space celestial body enhancement technology

Typical photos of deep space objects include both deep space objects and many stars. Of course, if you shoot globular clusters or open clusters, there are almost only stars.

There are some problems with the photos of stars. Stars are point sources, but due to the influence of the earth's atmosphere, the stars in the photos are all round spots. The diameter of the circular spot is proportional to the brightness of the star, and the edge is not very clear.

The difficulty lies in how to keep the stars unaffected when you deal with deep space objects, such as nebulae or galaxies. USM sharpening is a common sharpening method, and its principle is to improve the contrast of bright and dark areas. But a black circle will be made around the star, and excessive sharpening will make the star look abrupt.

One way is to select the stars, then reverse them and deal with other areas. Will it be difficult for every star to choose? Don't be afraid, there are several ways to do it, which is not as difficult as you think. The easiest way is to use the color range selection tool, and then select the middle of a bright star, which should be white. Changing the color tolerance can change the size of the selection range. Click OK. Are all the stars selected?

Then, the selection is expanded (select-modify-expand) by 2-5 pixels, and then feathered by 2 pixels (select-adjust edge-feather). If you reverse the selection, you can sharpen your deep space objects.

After choosing a star, there may be a situation. For example, most stars are overexposed and the center is white. Look carefully around the stars, there are colors, which are the original colors of the stars. If there is, copy the original layer, select the stars in the same way, and then blur the upper layer, and the surrounding colors will spread among the stars. Set the blending mode of the upper layer to "color" and the saturation to high, and the original color of the star can be restored. When doing this, make sure that the stars are always selected.

Layer mask

The best example is the famous Orion Nebula, which is a celestial body with a wide range of brightness. Its center is bright and its surroundings are dark. If you expose the center correctly, the surrounding structure will not be photographed at all. On the contrary, when shooting the surrounding structures, the center is overexposed and saturated.

One solution is to expose it twice, once to make the center exposure correct and once to make the periphery exposure correct. Synthesized by layer mask technology. The details of the center and periphery of the nebula can be perfectly displayed.

The first step is to assume that you have two photos of the correct exposure center and periphery. Paste two photos into two layers.

Step 2, take the one with correct peripheral exposure as the upper layer. Adjust the position so that the two layers are aligned.

Step 3, select the overexposed area in the upper center. You can probably choose it, and you don't need to find the overexposed edge too accurately.

Step 4: Hold down the alt key and click the "Add layer mask" icon to create a layer mask for the upper level.

Step 5, click on the mask displayed in the layer, and make a strong Gaussian blur to make the edge change gradually.

Step 6, click on the layer to preview the whole picture, and adjust the color levels and curves of the two layers to make the image look natural.