Camera interpolation, what is it? What is the interpolation of the camera in the telephone and what is it for? Schob not put zayvih ask.

Camera interpolation, what is it?

1. Type 8 Mp matrix, and 13 Mp image itself
2. You don't need to spin the darts to the matrix, the megapixels blow up right in the process.
3. If the pixel is split into sprats, then the squares will increase the size of the image. Do not give real permission. Smear the drawing.
4. interpolation - the value of the unknown value behind the given values.
   The degree of interpolation of a photograph (closer to the original) is inherent in a well-organized software
5. The camera sensor is 8MP, and the images are stretched up to 13MP. Turn it on unequivocally. Photos will be 13mp, but for clarity, like 8mp (digital noise will be more).
6. There is a real difference in the building there at the lines by mm without slacking off at any drop by 2mp.
7. Well, just blown up pixels
   For example, there are a lot of web cameras, it says 720 and so on.
8. Interpolation - in the case of a sensible sense - quotation when calculating the least folding function with the method of reaching the result, as close as possible to the absolute, dosjazhny, or for the help of the most accurate and correct ones.
   In this variant - it's just apparent, the programmers praise themselves for the fact that the signs of the phone are trochetted in the broken foldable outbuildings - cameras.

1. What matrices are better than Live MOS or CMOS ??? "Live MOS matrix is ​​traded for a variety of light-sensitive matrices that Panasonic expands and are also used for Leica cameras...
2. What is the Fresnel lens? 1. Fresnel lens2. Zvichayna lens The main advantage of the Fresnel lens is ...
3. Tell me, is the Fujifilm FinePix S4300 camera, 26-x ZOOM, like professional? Є slipped milnitsa milnitsa, supurzum. unsuitable for photo shoots. marvel here http://torg.mail.ru/digitalphoto/all/?param280=1712,1711amp;price=22000,100000 Mlynets, these are great...
4. How does a mirror video monitor look like an optical one? what is shorter? D-mirror vidoshukach - vіzuvannya vіdbuvaєsya behind the help of the system of mirrors, it's light to pass through the lens itself without a middle and...
5. Why is there a difference between CMOS sensors and CCD sensors for CCDs in video cameras? CMOS-matrix (CMOS) is a digital attachment, which can be mounted on one chip with the other guts.

Image interpolation is seen in all digital photographs at the singing stage, whether it is dematrization or scaling. You will see it quickly, if you change the size or the size of the image from one grid of pixels to another. Changing the size of the image is necessary, if you need to increase or change the number of pixels, then you can change the position in different ways: correcting the lens, changing the perspective, or rotating the image.

Need to change the size of the world, or the spreads are given by the same image, the results can be significantly different in the fallacy of the interpolation algorithm. Oscilki be-like interpolation is less close, images will be worn out like skin once, if interpolation is given. Tsya head poklikana to ensure better understanding of what is invested in the result, and they themselves will help you to minimize whether or not to spend the quality of the image, to interpolate.

concept

The essence of interpolation is based on the selection of obvious data for estimating the values ​​from unknown points. For example, if you wanted to know what the temperature of the day was, and how they fought її about the 11th year, you can let in the її meaning, zastosuvav linear interpolation:

Yakby, you have an additional temperature about 12 times, you could remember that before noon the temperature grew faster, and beat the additional temperature for quadratic interpolation:

The more you control the temperature around noon, the more complex (and more accurate) your interpolation algorithm can be.

An example of changing the size of the image

The interpolation of the image works for two images and it is possible to reach the best approximation for the color and brightness of the pixel, basing on the values ​​of the most important pixels. The stepping butt illustrates the scaling robot:

On the vіdminu vіd vіd kіlіvanі temіtrіvіv vіtrya vіshchena іnіvіdіv іdealnogo gradіentа, pіkselіv znachenya can zmіnjuvati na richly rіkіshe vіd dot to dot. As in the case of temperature, the more you know about the necessary pixels, the more quickly you can interpolate. Because of this, the results are quickly getting better in the world of image expansion, moreover, interpolation is by no means able to add detail to the image, which is in no way possible.

image wrapping example

The interpolation is also seen every now and then, if you rotate or change the perspective of the image. The front butt of the letter is deceptive, the shards are cremated, in which the interpolator sounds wrongly. The stepping butt is showing, as if it were possible, the detail of the image was lost:

Rotate by 90 ° do not make an insertion, it is not necessary to place the shards of the pixel on the border between the two (and as a reminder to add). Respect, as more details are ruined during the first turn, and as if they continue to fall during the next ones. Tse means what next unikati wrap, it is possible; yakscho nerіvno stavleniya frame vmagaє rotation, do not twist it more than once.

Introduce more results to win the so-called “bіkubіchny” algorithm and show the essence of the increase in brightness. To increase respect, as the bright contrast decreases through the decrease in the intensity of the color, as the dark halo vibrates a little light blue. The results can be significantly reducing the fallow in the interpolation algorithm of the imaged object.

Types of algorithms in interpolation

Globally accepted interpolation algorithms can be divided into two categories: adaptive and non-adaptive. Adaptive methods change stalely depending on the subject of interpolation (sharp borders, smooth texture), as well as non-adaptive methods process all pixels however.

Non-adaptive algorithms include: nearest sine method, bilinear, bicubic, spline, cardinal sine (sinc) function, Lanczos method and others. Fallow due to the folding of the stench of the vicorist, from 0 to 256 (or more) total pixels for interpolation. The more total pixels of the stench include, the more accurate they can appear, the more you can reach a significant increase in the hour of processing. Qi algorithms can be tweaked as for flaming, so scaling the image.

Adaptive algorithms include a lot of commercial algorithms in licensed programs such as Qimage, PhotoZoom Pro, Genuine Fractals and others. Many of them develop different versions of their algorithms (based on pop-pixel analysis), if they reveal the presence of the cordon - with the method of minimizing unacceptable interpolation defects in the areas, they are obvious. These algorithms are developed in the first place for maximization of defect-free detail of larger images, so for wrapping them, or changing the perspective of the image is not suitable.

Nearest Susid Method

The most basic of the most used interpolation algorithms, which will require the least hour of processing, only one pixel is the closest to the interpolation point. As a result, the skin pixel just becomes bigger.

Biliniyna interpolation

Biliniyna interpolation looks at a 2x2 square of pixels, which makes you feel invisible. As the interpolation of the values ​​of the victors is important, the averaging of these chotyroh pixels. As a result, the images look significantly smoother, lower than the result of the work of the nearest suside method.

The diagram of evil is seen to the point, if all the given pixels are equal, so the value is interpolated simply by its sum, divided by 4.

Bikubichna interpolation

Bіkubіchna іnterpolаtsіya ide one krok dіlі for bіlіnіyna, gazing at an array of 4x4 navkolyshnіh pіkselіv - 16 in total. Bіkubіchna іnterpolatsija rob significantly razkіshi izobrazhennya, nizh in front of two methods, і, perhaps, є optimal for spіvvіdshennyam hour of processing and brightness at the exit. For some reason, it has become standard for many image editing programs (including Adobe Photoshop), printer drivers, and camera interpolation.

Higher order interpolation: splines and sinc

There are a lot of other interpolators, if you take more than the necessary pixels to the point of respect, and in this way you will require more intensive calculations. The algorithms include splines and cardinal sine (sinc), and they save more information about the image after the interpolation. As a last resort, the stench is exclusively brown, if the image shows a lot of turns that change the perspective beyond the edges. However, for one-time improvements or rotations, such high-order algorithms give slightly visual improvement with a significant increase in processing time. Moreover, in some cases, the algorithm of the cardinal sine on a smooth distance is higher, lower bicubic interpolation.

Interpolation defects that are guarded.

The use of non-adaptive interpolators is induced to achieve the optimal balance between three minor defects: boundary halo, rosette and stage part.

To develop the most variant of non-adaptive interpolators and to change the error, or to change one of the defects that have been introduced for the other two - as a legacy, as at least one of them will be remembered. Respectfully, the borderline halo is similar to a defect, which is born to increase the sharpness for an additional unsharp mask, and as if it raises the sharpness, which is given, for the additional strength of clarity.

Adaptive interpolators may or may not create the above-described defects, but stinks can also generate an unpowered visual image of the texture or single pixels on a large scale:

On the other hand, the “defects” of the adaptive interpolators can also be seen as an advantage. Shards of eyes can be seen in areas with a fine texture, such as leaves, details right down to the smallest details, like little ones can fool the eye on the wall (for the most beautiful sights of the material).

Zglazhuvannya

Smoothing or anti-aliasing is a process that helps to minimize the appearance of steps or jagged diagonal lines, like giving the text or images a rough digital look:

	300%

Gladzhuvannya saw and descended and created the enemy of softness between that high and varied building. I take it to the point of respect, the ideal borders intersect the sum of the pixels. The step of the boundary is simply rounded up or down without an intermediate value, even as the boundary is smoothed out, in proportion to the fact that a rich cordon was squandered into a leather pixel:

Important for the improvement of the image is the loss of the transcendental step frequency as a result of the interpolation. A lot of adaptive interpolators signify the presence of cordons and are adjusted with the method of minimizing the stage part, saving at the same sharpness of the cordon. Oskіlki smoothed out the boundary between information about their camp with a larger high distribution of buildings, as much as possible, a tighter adaptive (original boundary) interpolator may want to often reconstruct the cordon for the larger ones.

Optical and digital zoom

A lot of compact digital cameras can be both optically and digitally enhanced (zoom). The optical zoom is controlled by the variability of the lens, so that the light shines on the digital sensor. In contrast, the digital zoom lowers the brightness, leaving the shards to simply interpolate the image after taking it out by the sensor.

optical zoom (10x)		digital zoom (10x)

Ignore those photos with digital zoom settings to compensate for the same number of pixels, and the detail is noticeably smaller, lower with optical zoom settings. Digital zoom, for a wink of vipadkіv, if Vіn helps to display in the distance of the object on the RK-screen of the camera. On the other hand, if you want to record in JPEG and want to crop and enlarge the image, digital scaling may be more important in the fact that this interpolation is built before the defect in the compression is introduced. If you see that you need digital zoom more often, buy a teleconverter, and even more shortly a lens with a large focal length.

Sensors - tse attachments, which indicate only gradations of gray (gradations of light intensity - from white to completely black). So that the camera could distinguish colors, silicon for the additional process of photolithography is superimposed with an array of color filters. In quiet sensors, microlenses are deviated, filters are placed between lenses and a photodetector. In scanners, trilinear CCDs are de-wired (there are three CCDs in order, which react differently to red, blue and green colors), or in high-end digital cameras, three sensors are also wicked, the skin sensor is filtered by the light of a songful color. (Respectfully, in some cameras with a number of sensors, combinations of several colors are used for filters, and not three standard ones). For attachments with one sensor, such as the most advanced digital cameras, for processing different colors, arrays of color filters (color filter arrays, CFA) are used.

In order to give the skin pixel its own main color, the filter of the main color is brushed over it. Photons, first of all, spend on a pixel, go through the filter, which skips less than the whiff of your color. The light of the day is just to be clayed with a filter. Vcheni vznachal, scho be-what color in the spectrum can be taken away from the mix of just a few of the main colors. The RGB model has three such colors.

For skin zastosuvannya arrays of color filters are rozroblyayutsya. In addition, in most sensors of digital cameras, the most popular are arrays of filters in the Bayer pattern. This technology was found in the 70s by the Kodak company, if the studies were carried out in the gallery of a spacious rose floor. In my system, the filters are arranged in a superior order, in a checkered order, and the number of green filters is larger, lower is red or blue. The order of stitching is such that red and blue filters are stitched between green.

Such a kіlkіsne spіvvіdnoshnja is explained by the everyday human eye - vin sensitive to the green light. And the checker order ensures the same colors of the image regardless of how you trim the camera (vertically or horizontally). When reading information from such a sensor, the colors are recorded sequentially in rows. The first row is to blame buti BGBGBG, the offensive - GRGRGR and so on. This technology is called sequential RGB (sequential RGB).

In PZZ cameras, all three signals are sent together not on the sensor, but in the attachment, the image is formed, even after that, as a signal of transformations from the analog type is digital. For CMOS sensors, confusion can be seen directly on the chip. In any case, the primary colors of the skin filter are mathematically interpolated with the improved colors of the susceptible filters. Respectfully, if there are more dots in the picture, they change the main colors, and it’s less effective to represent a pure red, blue or green color.

For example, in order to designate, having added secondary pixels to the color of the central one, with linear interpolation, a matrix of pixels with a size of 3x3 is processed. Take, for example, the simplest way - three pixels - with blue, red and blue filters, arranged in one row (BRB). It is acceptable that you try to take the resulting value of the color of the red pixel. As all colors are equal, the colors of the central pixel are mathematically calculated as two parts of blue to one part of red. Really, the algorithms to create simple linear interpolation are richly collapsible, the stench is to protect the values ​​of all the necessary pixels. If the interpolation looks bad, then there will be teeth on the border of the color change (otherwise the color artifacts appear).

Significantly, the word "dozvіl" in the sphere of digital graphics is not correct. Purists (otherwise pedantry - to whom it is more appropriate), know about photography and optics, know that the building is different - the world of the building of the human eye, or the device to distinguish the okrem lines on the page of the allowable, for example, on the page below the ISO, indications. But in the computer industry, it is accepted that it is allowed to name a number of pixels, and as it happened, we also pіdemo tsієї convention. Even navit retailers call the retail building a number of pixels at the sensor.

Shall we?

Expand the image file to be stored as number of pixels (admitted). The bigger the pixels, the bigger the file. For example, an image sensor in the VGA standard (640x480 or 307200 active pixels) takes about 900 kilobytes in a squeezed look. (307200 pixels by 3 bytes (R-G-B) = 921600 bytes, which is approximately 900 kilobytes) A ​​16 MP sensor image takes approximately 48 megabytes.

It would have been possible to do something like this - to check the number of pixels on the sensor, to determine the size of the image, which to go out. Tim is no less, the chambers of the chambers represent a bunch of different figures, and it is hard to say that it is the right thing to do in the chamber.

The total number of pixels includes all pixels that are physically present in the sensor. Ale, active vvazhayutsya tі, yakі take the fate of the original image. Close to five hundred pixels of the required pixels do not take part in the original image. These are defective pixels or pixels that are picked up by the camera for other indications. For example, you can use masks to define the dark strum or to define the frame format.

The frame format is the ratio between the width and the height of the sensor. For some sensors, for example, with a wide range of 640x480, the speed is 1.34: 1, which matches the frame format of more computer monitors. This means that the images created by such sensors are accurately inserted into the monitor screen, without forward framing. For bagatioh devices, the frame format is similar to the traditional 35-millimeter swimming pool, the ratio is equal to 1:1.5. Tse allows you to work the images of the standard size and form.

Interpolation allowed

Crim of optical permission (the ability of pixels to react to photons is real), it also allowed, an increase in the software and hardware complex, for additional algorithms that interpolate. As an interpolation of colors, in the interpolation it is possible to mathematically analyze the data of the adjacent pixels. With whichever interpolation occurs, intermediate values ​​are created. Such a "winning" of new data can be carried out smoothly, with which interpolation the data will be between average, between real optical data. But sometimes for such an operation, they can be blamed for various changes, artifacts, appear to be created, in which the quality of the image is less lost. That's a lot of pessimism in vvazhayut, scho interpolation allowed - not a way to improve the quality of the image, but rather a way to increase the files. When choosing, I will add respect, as a distribution of buildings is indicated. Not varto for the sake of high interpolation of different buildings. (It is denoted as interpolated or enhanced).

Another process of image processing on a software level is sub-sampling. In essence, the process of wrapping up to interpolation. This process is carried out at the stage of image processing, even after the transformation from an analog digital look. When you see the data of different pixels. For CMOS sensors, this operation can be carried out on the chip itself, by turning on the reading of the first rows of pixels, or reading the data only from the selected pixels.

Subsampling wins two functions. First, for the strengthening of the tribute - to save more znіmkіv from the memory of the singing rose. The smaller the number of pixels, the smaller the amount of time the file is expanded, and the more you can fit on the memory card, or at the internal memory I will add, and the easier it is for you to capture photos on the computer or memory cards.

Another function of this process is the creation of an image of a song expansion of song goals. Cameras with a 2MP sensor are tough enough to take a picture of a standard photo with a size of 8x10 inches. Alternately, you can try to send such a photograph by mail, it’s worth remembering to increase the size of the sheet. Subsampling allows you to process the image in such a way that it looks normal on the monitors of your friends (so you don’t put it as a meta detail) and when you do it, it was possible to do it on machines with the correct data.

Now, if we are familiar with the principles of robotic sensors, we know how to get an image, let's take a look at a little more clearly and deal with more complex situations that are blamed for digital photography.

What is camera interpolation?

All current smartphones have built-in cameras that allow you to capture images more with the help of special algorithms. From the mathematical point of view, the interpolation is the method of revealing the intermediate values ​​of the number behind the apparent set of discrete parameters.

The effect of interpolation seems to be guessing foolish. Smartphone software does not increase the clarity and sharpness of the image. It just expands the picture to the required size. Some manufacturers of smartphones write on the packaging of their products that the camera is built in, it can reach up to 21 megapixels. It is best to go about the interpolation of the image itself, as it may be low in quality.

Types of interpolation

Nearest Susid Method

The method is considered basic and belongs to the category of the simplest algorithms. Pixel parameters are determined based on one nearest point. As a result of mathematical expansions, the expansion of the skin pixel increases twice. Using the nearest pixel method does not require great computational efforts.

Biliniyna interpolation

The pixel value is determined based on the data about the nearest points captured by the camera. As a result, the calculation becomes more important for the averaged parameters of 4 pixels, so as to locate the exit point. Bilinear interpolation allows you to smooth transitions between the color boundaries of objects. Images, otrimani z vikoristannym tsgogo method, significantly change the quality of the picture, interpolation by the closest pixel method.

Bikubichna interpolation

The value of the color of the point that is being searched is calculated based on the parameters of the 16 nearest pixels. Krapki, yakі roztashovanі nayblizhche, otrimuyut maximum vag during rozrahunka. Bicubic interpolation actively wins over to the software security of today's smartphones and allows you to capture a clear image. Zastosuvannya method due to the significant tightness of the central processor and the high space of the built-in chamber.

Shchob do not put zayvih pitan:

Pluses and minuses

In science fiction films, it is often shown how the camera captures the appearance of the transient and transmits digital information to the computer. The machine captures the image, recognizes the photograph and recognizes the person in the database. In real life, interpolation does not add new details to the image. Vaughn just zbіshuє vyhіdnu picture for the help of a mathematical algorithm, improving its quality to a pleasant level.

Interpolation defects

The most common defects that are blamed for image scaling are:

• Step-part;
• rosemity;
• Halo effect (halo).

All interpolation algorithms make it possible to achieve a perfect balance of defect overshoots. The change in the frequency step of the obov'yazkovo was more pronounced than the increase in the brightness of the image and the appearance of a halo. Strengthening the sharpness of the image will bring the picture to the height of the sharpness of the picture. A number of interpolation defects can be identified by different graphical "noises", which can be predicted with the maximum image enhancement. To talk about the appearance of “vipatic” pixels and textures that are not powerful for a given object.