JP2011250260A - Image processing apparatus and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing apparatus that suppresses excess in contrast and chroma without change of the color temperature, chroma, and hue after compression of a dynamic range.SOLUTION: An image processing apparatus has: a signal supply part that inputs a luminance signal and first and second color-difference signals; a dynamic range compression part that compresses the dynamic range of the inputted luminance signal, and outputs the luminance signal whose dynamic range is compressed; and a multiplication part that multiplies the first and second color-difference signals by a coefficient depending on a ratio of the luminance signal whose dynamic range is compressed to the inputted luminance signal.

Description

  The present invention relates to an image processing apparatus and an image processing method for compressing a wide dynamic range image signal photographed by a camera.

  2. Description of the Related Art Conventionally, in-vehicle cameras and security surveillance cameras have been widely used. For example, a camera is mounted on a vehicle, and a video around the vehicle is captured and displayed on a display to assist a driver. By the way, the imaging environment of a vehicle-mounted camera is from a clear sky to a dark, and a wide dynamic range is required. For example, assuming an illuminance of 0.15 to 150,000 lx, the dynamic range is 120 dB, and a 20-bit data width is required. When an image taken by a camera is subjected to compression processing, a non-linear circuit is generally introduced in the photoelectric conversion unit of the image sensor so that the optical input 120 dB is compressed and a signal of about 10 to 12 bits is output.

  On the other hand, the dynamic range of an in-vehicle display device that displays an image taken by a camera is about 36 dB to 48 dB (6 to 8 bits). In order to display an image signal with a wide dynamic range on the display device, the dynamic range is further increased. Range compression is required. However, simply compressing 10 to 12 bits to 6 to 8 bits results in an image with low contrast and poor visibility.

  As a conventional dynamic range compression technique, an example using a Retinex model as described in Non-Patent Document 1 is well known. Non-Patent Document 1 discloses an image processing technique for automatically converting the brightness and contrast of a captured image so that it can be easily seen using a Retinex model considering visual characteristics and a method for creating a mappin curve suitable for each region of an image. ing.

  In this Non-Patent Document 1, a component extraction unit that extracts a reflectance component and an illumination bright light component, an intensity conversion unit that converts the intensity of the reflectance component, and a mapping curve suitable for each region by dividing the illumination light component into regions. And a tone mapping unit that performs tone mapping.

  Patent Document 1 discloses an image processing apparatus that compresses the dynamic range of an input image. In the example of Patent Document 1, an extraction unit that extracts an illumination component and a reflectance component from an input image, a compression unit that compresses the dynamic range of the extracted illumination component, a compressed illumination component, and an extracted reflectance, respectively. And a synthesis means for synthesizing the components.

  In Patent Document 1, similarly, processing based on periodic downsampling is performed hierarchically from photoelectric conversion signals having pixel (RGB) arrays having periodically different spectral sensitivities such as a Bayer array, and the illumination component L Estimating. That is, three different signals (RGB) are each subjected to independent processing, the illumination component for each of RGB is estimated, the reflectance component of each RGB is obtained using this illumination component, and each illumination component is compressed. The compressed R′G′B ′ is obtained by the product with the respective reflectance components.

  However, in the conventional compression technique described above, since each RGB signal is subjected to nonlinear processing, there is no guarantee that the relative level of the output RGB signal is always the same as the relative level of the input RGB signal. For this reason, depending on the content of the image, the color temperature and saturation / hue may differ between input and output. In the compression based on the Retinex model, the illumination component is compressed, so that the contribution of the reflection component is relatively large. If the compression ratio is set large, the contrast and saturation may be excessive.

Toshiba Review Vol.64, No.6 (2009) "ContrastMagic TM, brightness optimization technology for captured images"

JP 2009-77274 A

  In the conventional compression technique, each RGB signal is subjected to nonlinear processing, so there is no guarantee that the relative level of the output RGB signal is always the same as the relative level of the input RGB signal, and depending on the content of the image, the color temperature and There are cases where saturation and hue differ between input and output. In the compression based on the Retinex model, the illumination component is compressed, so that the contribution of the reflection component becomes relatively large, and the contrast and saturation may be excessive.

  In view of such circumstances, the present invention provides an image processing apparatus and an image processing method that suppresses excessive contrast and saturation without changing color temperature, saturation, and hue after compression of a dynamic range. With the goal.

  An image processing apparatus according to an embodiment of the present invention includes a signal supply unit that inputs a luminance signal and first and second color difference signals, and dynamic range compression that compresses and outputs the dynamic range of the input luminance signal. And a multiplier that multiplies the first and second color difference signals by a coefficient corresponding to a ratio between the luminance signal with the dynamic range compressed and the input luminance signal, and converts the coefficient. The luminance signal and the color difference signal thus output are output.

  An image processing apparatus according to an embodiment of the present invention includes a signal supply unit that inputs a luminance signal and first and second color difference signals, and dynamic range compression that compresses and outputs the dynamic range of the input luminance signal. And a first multiplication that multiplies the first and second color difference signals by a first coefficient α corresponding to a ratio between the luminance signal with the dynamic range compressed and the input luminance signal, and outputs the first and second color difference signals. And a specific color area, and when the first and second color difference signals exist in the specific color area, the first coefficient α is multiplied by a second coefficient β (> 1). And a second multiplication unit for increasing and outputting the color difference signal.

  The image processing method according to an embodiment of the present invention inputs a luminance signal and first and second color difference signals, compresses and outputs the dynamic range of the input luminance signal, and the dynamic range is compressed. The first and second color difference signals are multiplied by a coefficient corresponding to the ratio between the input luminance signal and the input luminance signal and output.

  According to the image processing apparatus of the embodiment of the present invention, the color temperature, the saturation, and the hue are not changed before and after the compression of the dynamic range, and excessive saturation is prevented after the compression of the dynamic range. Can do. Further, it is possible to emphasize and output the saturation of a signal in a preset color region.

1 is a block diagram showing the overall configuration of an image processing apparatus according to an embodiment of the present invention. The block diagram which shows an example of a dynamic range compression part. The characteristic view which shows the input-output characteristic of the dynamic range compression part of FIG. FIG. 5 is a chromaticity diagram illustrating a specific color region set by a color determination unit.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of an image processing apparatus according to an embodiment of the present invention. 1, the image processing apparatus 10 compresses the dynamic range of the luminance signal Y and the matrix circuit 11 that converts the RGB signal obtained by photographing with the camera 20 into the luminance signal (Y) and the color difference signals (U, V). A dynamic range compression unit 12 (hereinafter simply referred to as a compression unit 12) is provided. The matrix circuit 11 constitutes a signal supply unit that inputs the luminance signal and the first and second color difference signals.

  Further, the calculation unit 13 that calculates the coefficient α determined by the ratio (Y ′ / Y) of the luminance signal Y before compression and the luminance signal Y ′ after compression, and the coefficient α (= Y ′ / Y calculated by the calculation unit 13). ) Is multiplied by the color difference signals U and V to provide output color difference signals U ′ and V ′. Further, when a specific color region (for example, red, yellow, blue) is preset by the color determination unit 17 and it is determined that the color difference signals U and V correspond to the specific color region, the color difference signals U and V are multiplied. A multiplication circuit 16 that multiplies the coefficient α by an appropriate coefficient β (> 1) to emphasize the saturation of a specific color region.

  A conversion circuit 18 that converts the compressed luminance signal Y ′ and the color difference signals U ′ and V ′ multiplied by a predetermined coefficient into RGB signals and obtains output R ′, G ′, and B ′ signals is provided. R ′, G ′, B ′ signals are supplied to the display device 30.

  In the embodiment of FIG. 1, a wide dynamic range camera output (RGB signal) is converted into a luminance signal Y and color difference signals U and V by the matrix circuit 11, and then the dynamic range of the luminance signal Y is compressed by the compression unit 12. To Y ′. Next, the coefficient α (= Y ′ / Y) is obtained by the calculation unit 13, and the color difference signals U ′ and V ′ are obtained by multiplying the color difference signals U and V by the coefficient α by the multiplication units 14 and 15. The Y ′, U ′, and V ′ signals may be hereinafter referred to as converted output.

  Since the converted outputs Y ′, U ′, and V ′ compress and convert the luminance signal Y into Y ′, the color temperature does not change even after compression. Further, the color difference signals U and V are respectively converted to the color difference signals U ′ and V ′ by multiplying by the same coefficient α (= Y ′ / Y), so that the dynamic range is compressed at the same rate as the luminance, and excessive saturation is not obtained. It is prevented and the hue does not change.

  FIG. 2 is a block diagram illustrating an example of the dynamic range compression unit 12. The dynamic range compression unit 12 compresses the dynamic range of the luminance signal Y by a method based on, for example, the Retinex model and converts it into a Y ′ signal. The illumination component detection unit 121, the compression unit 122, and the reflectance component detection unit 123 and a multiplication circuit 124.

  According to Retinex, the input pixel value I can be modeled by the product of the illumination light component L and the reflectance component R, that is, I = L × R. The illumination light component L is a component of illumination light irradiated on the subject, and the reflectance component R is an image component of the subject.

  The illumination component detection unit 121 detects the illumination light component L for each pixel from the input pixel value I by, for example, filter processing using a low-pass filter. The compression unit 122 compresses the dynamic range of the illumination component detected by the illumination component detection unit 121 with a preset compression characteristic. For example, the inclination of the output illumination component L ′ becomes gentle as the illumination component L increases. Has parabolic properties.

  The reflectance component detection unit 123 detects the reflectance component R for each pixel from the input pixel value I by filter processing using, for example, a high-pass filter. The multiplication circuit 124 multiplies the illumination component L ′ and the reflectance component R with a compressed dynamic range, and generates an output pixel value I ′ = L ′ × R.

  FIG. 3 is a characteristic diagram showing input / output characteristics of the dynamic range compressor 12 of FIG. 3A shows the input illumination component and reflectance component, and FIG. 3B shows the output illumination component and reflectance component. As can be seen from FIG. 3, the compression of the dynamic range is performed on the illumination component, and the reflectance component is not compressed. Accordingly, it is possible to prevent the visibility from being deteriorated due to a decrease in contrast.

  1 determines a specific color area (for example, red, yellow, and blue) in advance, and determines that the color difference signals U and V exist in the specific color area, the multiplication circuit 16 In this case, the coefficient α multiplied by the color difference signals U and V is multiplied by an appropriate coefficient β (> 1) to enhance the saturation of a specific color region for emphasis.

  For example, when a video around a vehicle is shot with a vehicle-mounted camera, the color area of an emergency vehicle warning light includes red, yellow, and blue components. Accordingly, when it is desired to accurately photograph the warning light, red, yellow, and blue color regions are set in advance, and when it is determined that the color difference signals U and V correspond to this color region, the coefficient α multiplied by the color difference signals U and V Is multiplied by a coefficient β (> 1) to increase the saturation of the color area of the warning light. As a result, the saturation of the color area of the warning light can be highlighted when displayed on the display 30.

  FIG. 4 is a chromaticity diagram illustrating a specific color region set by the color determination unit 17. In FIG. 4, a triangle indicated by RGB is a color range that can appear on the display 30, and the color region set by the color determination unit 17 is indicated by three circles (red, yellow, and blue). The color region set by the color determination unit 17 is not limited to red, yellow, and blue, and other color regions can be set.

  According to the embodiment described above, the luminance signal Y is compressed and converted into Y ′, so that the level is merely changed, and the color temperature does not change even after compression. In addition, since the color difference signals U and V are multiplied by the same coefficient α (= Y ′ / Y) and compressed to the color difference signals U ′ and V ′, the relative levels of Y ′, U ′, and V ′ do not change and dynamic Saturation and hue do not change before and after range compression. Further, since the color difference signals U and V are compressed at the same level as the luminance, it is possible to prevent excessive saturation after compression. In addition, the saturation of the signal of a preset color area can be emphasized and output, and a specific color area, for example, a warning light of an emergency vehicle can be emphasized and displayed. It is effective when displaying.

  Note that the image processing apparatus according to the embodiment described above can be applied not only to a vehicle but also to a case where an image obtained by a monitoring camera or the like is processed, and the embodiment of the present invention includes the scope of claims. Various modifications are possible without departing from the scope.

DESCRIPTION OF SYMBOLS 10 ... Image processing apparatus 11 ... Matrix circuit 12 ... Dynamic range compression part 13 ... Calculation part 14, 15 ... Coefficient alpha multiplication part 16 ... Coefficient beta multiplication part 17 ... Color determination part 18 ... Conversion part 20 ... Camera 30 ... Display 121 ... Illumination component detector 122 ... Compressor 123 ... Reflectance component detector 124 ... Multiplier

Claims (7)

A signal supply unit for inputting the luminance signal and the first and second color difference signals;
A dynamic range compression unit that compresses and outputs the dynamic range of the input luminance signal;
A multiplier for multiplying and outputting the first and second color difference signals by a coefficient corresponding to a ratio between the luminance signal with the dynamic range compressed and the input luminance signal;
An image processing apparatus that outputs a converted luminance signal and color difference signal. The signal supply unit includes a matrix circuit that converts an RGB signal obtained by photographing with a camera into a luminance signal Y and first and second color difference signals U and V;
The dynamic range compression unit outputs a luminance signal Y ′ obtained by compressing the dynamic range of the luminance signal Y,
The image processing apparatus according to claim 1, wherein the multiplication unit multiplies the first and second color difference signals U and V by the coefficient (α = Y ′ / Y) and outputs the result.   5. A conversion circuit for converting the luminance signal Y ′ having the compressed dynamic range and the first and second color difference signals U ′ and V ′ multiplied by the coefficient α into RGB. 2. The image processing apparatus according to 2. A signal supply unit for inputting the luminance signal and the first and second color difference signals;
A dynamic range compression unit that compresses and outputs the dynamic range of the input luminance signal;
A first multiplier that multiplies the first and second color difference signals by a first coefficient α corresponding to a ratio between the luminance signal with the dynamic range compressed and the input luminance signal;
When a specific color area is set and the first and second color difference signals exist in the specific color area, the first coefficient α is multiplied by a second coefficient β (> 1) to obtain a color difference. An image processing apparatus comprising: a second multiplication unit that increases and outputs a signal. Input the luminance signal and the first and second color difference signals,
Compress and output the dynamic range of the input luminance signal,
An image processing method comprising: multiplying the first and second color difference signals by a coefficient corresponding to a ratio between a luminance signal with a compressed dynamic range and the input luminance signal, and outputting the result.   6. The image processing method according to claim 5, wherein the luminance signal having the compressed dynamic range and the first and second color difference signals multiplied by the coefficient are converted into RGB and supplied to the display.   When a specific color area is set and the first and second color difference signals exist in the specific color area, the coefficient is multiplied by a second coefficient (> 1) to increase the color difference signal. 6. The image processing method according to claim 5, wherein the image processing method outputs the image.

Images