TWI587245B - Image enhancement method - Google Patents

Description

Image enhancement method

The present invention relates to an image enhancement method; more particularly, the present invention relates to an image enhancement method that increases the dynamic range of a high contrast image so that high contrast image detail can be clearly presented on a low contrast display.

Based on the increasing demand for vision by humans, current display devices have evolved toward high dynamic range and high resolution. Generally static or moving images are intended to clearly show details, which are highly correlated with the resolution, contrast and dynamic range of the image. At present, the software and hardware technology has been continuously improved, and the resolution of images has been continuously improved. It has been upgraded from 360p, 480p, 720p, 1080p to the now popular 2K TV, or even the mainstream 4K and 8K TVs in the future. In addition to resolution, dynamic range is also an important factor in the clarity of image detail. High-end TVs now have a high dynamic range (HDR) image adjustment capability as a selling point. Generally speaking, a high dynamic range image refers to a high brightness difference between the bright portion and the dark portion of the image, and the image detail is much larger than the Low Dynamic Range (LDR) image. In other words, the greater the difference in brightness between the brightest white and the darkest black of the image, the more clearly the image detail can be displayed, which brings the image closer to the image viewed by the real human eye. Need to know only to improve the resolution, but if If the dynamic range is not improved, the contribution to the image detail is limited, and the image contrast is improved. Although the brightness of the image is increased, the appearance of the details of the image and the dark portion is still limited. Therefore, the increase in dynamic range plays a key role in the presentation of image detail. High-motion images not only increase the brightness, but also show more details in the dark and high light. The images are more vivid, the colors are more full and true, and the color gradient is more delicate, so it can be more natural and real. image. As the size of the TV zooms in, the details of the image will be examined more closely, so the technique of increasing the dynamic range is also more important.

At present, a general computer monitor or a television has a low dynamic range of images, so that high-contrast images are limited in display thereon, and the image details of the bright and dark portions of the image cannot be clearly displayed. Due to the large size of the high contrast image, this often results in resource consumption. Although high-end TVs emphasize the dynamic range of their adjustable images, the cost is still too high and the popularity is not easy.

Based on the above, it is still necessary to develop a method that can greatly improve the dynamic range of the image, so that the details of the image and the dark portion can be more clearly presented, so as to meet the future demand for high-size, high-quality images.

The invention provides an image enhancement method, in particular, the dynamic range of the high contrast color image can be further enhanced, so that the image details of the bright part and the dark part can be more clearly presented, so that the high contrast color image can be used in various specifications. The display is characterized by its high quality requirements for images.

In an embodiment, the present invention provides an image enhancement method. The method comprises: obtaining an HSV color space of one color image; taking out the color image corresponding to one H channel image, one S channel image and one V channel image of the HSV color space; separating the V channel image by using a bidirectional filter A base layer image and a detail image are obtained; the base layer image is processed by using a limited contrast adaptive histogram equalization step to obtain a contrast enhanced base layer image; and the contrast enhanced base layer image and the detail image are combined to obtain a new V channel. Image; and combining the new V channel image with the original H channel image and the S channel image to obtain a new color image after the original color image is enhanced and contrasted.

In the above image enhancement method, the detail image is obtained by subtracting the base layer image from the V channel image.

In the above image enhancement method, the bidirectional filter can be represented by the following relationship: ;as well as Where J _s is the result of one pixel s processed by the bidirectional filter, I _p and I _s are the intensity values of one pixel p and pixel s, respectively, Ω is the whole image, and f and g are respectively in a spatial domain and one The Gaussian smoothing function of the intensity domain, K _s is a normalization function.

In the image enhancement method, the contrast-adapting histogram equalization step is to crop the histogram formed by the pixels of the base image by a predetermined threshold to limit the cumulative distribution function value (CDF) amplification range, histogram The value being cropped can be expressed in the following relationship: Where M x N is the total pixel value of the base layer image, L is the maximum image intensity value, α is a clipping factor, and S _Max is the maximum slope.

In the above image enhancement method, the color image is presented on a display, and the contrast of the color image is higher than the contrast exhibited by the display.

S101~S106‧‧‧Steps

1 is a schematic flow chart of an image enhancement method according to an embodiment of the present invention; FIG. 2 is a high contrast color image displayed on a low contrast display; and FIG. 3 is a second figure; The high-contrast color image is taken out of the V-channel image; the fourth picture shows the base layer image obtained by processing the V-channel image of FIG. 3 through the bidirectional filter; and the fifth figure shows the V-channel of FIG. The detailed image obtained by the image processed by the bidirectional filter; the sixth figure shows the new base layer image after the base layer image of FIG. 4 is enhanced by the restriction contrast adaptive histogram equalization step; FIG. 7 shows a new V-channel image combined with the original detail image in Figure 6; Figure 8 is a diagram showing a new color image in which the new V channel image in Fig. 7 is combined with the original H channel image and the S channel image.

Hereinafter, a plurality of embodiments of the present invention will be described with reference to the drawings. For the sake of clarity, many practical details will be explained in the following description. However, it should be understood that these practical details are not intended to limit the invention. That is, in some embodiments of the invention, these practical details are not necessary. In addition, some of the conventional structures and elements are shown in the drawings in a simplified schematic manner in order to simplify the drawings.

Please refer to FIG. 1 , which is a flow chart of an image enhancement method according to an embodiment of the invention. The image enhancement method proposed by the present invention generally comprises the following steps.

Step S101: Acquire one of the color images of the HSV color space.

Step S102: The color image is taken out to correspond to one H channel image, one S channel image and one V channel image in the HSV color space.

Step S103: separating the V channel image into a base layer image and a detail image by using a bidirectional filter.

Step S104: processing the base layer image by using a limited contrast adaptive histogram equalization step, thereby obtaining a new base layer image with contrast enhancement.

Step S105: Combining the contrast enhanced new base layer image with the detail image to obtain a new V channel image.

Step S106: Combining the new V channel image with the original H channel image and the S channel image to obtain a new color image after the original color image is enhanced and contrasted.

The implementation details of the steps are described below. Generally, color images are processed, and a RGB color space is generally used to describe a color image; that is, a color space composed of three primary colors of red, green, and blue. However, directly enhancing the RGB three primary colors will result in undesirable hue and saturation distortion. In order to faithfully preserve the hue and saturation of the original input image, in step S101, the RGB color space is converted to an HSV color space more in line with human visual characteristics. HSV color space, which refers to hue, saturation and lightness (Hue, Saturation, Value). Hue (H) refers to the basic properties of color, that is, the commonly known color names, such as red, yellow, and so on. Saturation (S) refers to the purity of the color. The higher the color, the softer the color. The lower the color, the more gray, which is represented by a value of 0-100%. Brightness (V) is the brightness, expressed as a 0-100% value. Therefore, in step S102, the H channel image, the S channel image, and the V channel image of the color image are taken out, and the V channel image is taken as the object of subsequent processing.

In order to adjust the contrast of the original image while retaining the image details, in step S103, the V channel image is separated into a base layer image and a detail layer image by using a bidirectional filter. . The bidirectional filter is a non-linear filter, which can be expressed as follows: ;as well as Where J _s is the result of one pixel s processed by the bidirectional filter, I _p and I _s are the intensity values of one pixel p and pixel s, respectively, Ω is the whole image, and f and g are respectively in a spatial domain and one The Gaussian smoothing function of the intensity domain, K _s is a normalization function.

Next, in step S104, the base layer image is processed using a limited contrast adaptive histogram equalization step. Histogram equalization (HE) is commonly used in the field of image processing and is a method for improving image contrast. The histogram equalization achieves the purpose of adjusting the image characteristics by quantizing the image color and performing statistical means. For example, a color image is considered to be composed of a plurality of pixels having different gray values, and is classified using a statistical technique. If the pixel gray value is unevenly distributed in the histogram statistics and is limited to a certain area, it means that the contrast of the overall image is very low. If the pixel gray value distribution of the histogram statistics is adjusted so that it spreads evenly throughout the distribution area, the contrast of the overall image will be enhanced. However, the traditional histogram equalization method has its limitations, which is not helpful for improving the dynamic range of high contrast color images, and may even cause image distortion due to excessive amplification of noise. In order to solve this problem, the present invention performs adjustment using a contrast limited adaptive histogram equalization (CLAHE).

The limit contrast adaptive histogram equalization step crops the histogram formed by the pixels constituting the base image with a predetermined threshold to limit the cumulative distribution function value (CDF) amplification amplitude. The value that the histogram is cropped can be expressed in the following relationship: Where M x N is the total pixel value of the base layer image, L is the maximum image intensity value, α is a clipping factor, and S _Max is the maximum slope.

After the step S104, the base layer image is processed to obtain a contrast-enhanced new base layer image. Next, in step S105, the new base layer image is merged with the original detail image to obtain a new V channel image. Then, in step S106, the new V channel image is merged with the original H channel image and the S channel image to obtain a new color image with high contrast and high dynamic range.

The effects of the above steps S101 to S106 are explained in Figs. 2 to 8. In Fig. 2, a high contrast image is displayed on a low contrast display. Based on the hardware limitation, it can be seen that the image details completely disappear, resulting in a low dynamic range and image distortion. In Fig. 3, the V-channel image in which the high contrast color image in Fig. 2 is represented in the HSV color space and taken out is shown. In Fig. 4 and Fig. 5, the base layer image and the detail image obtained by processing the V channel image of Fig. 3 through the bidirectional filter are respectively shown. The detail image in FIG. 5 is obtained by subtracting the base layer image in FIG. 4 from the V channel image in FIG. In Fig. 6, the new base layer image obtained by the enhancement of the contrast contrast adaptive histogram equalization step of the base layer image of Fig. 4 is shown. It can be seen that the contrast and image detail of the new base layer image have been greatly improved. In Figure 7, the combination of the new base image in Figure 6 and the original detail image is shown. After the new V channel image, the new V channel image has higher contrast and image detail than the original V channel image, and the dynamic range of the displayed image has been improved. In Fig. 8, a new color image in which the new V channel image in Fig. 7 is combined with the original H channel image and the S channel image is shown. From Fig. 8, it can be seen that the new color image has higher contrast and image detail than the color image in the original Fig. 2. In other words, on the low-contrast display of the same specification, after the original color image is processed by the image enhancement method of the present invention, more details of the bright and dark portions of the image can be seen, that is, the dynamic range of the original color image has been greatly improved. .

The image enhancement method proposed by the present invention can be well applied to a large-sized, high-definition television to be developed in the future. Based on the progress of hardware, the current TV has not only have the function of displaying images, but has developed towards multi-functional smart TV. Therefore, a computer device having an arithmetic function is provided on the television, for example, a processor having a logic operation function and a related output port. The image enhancement method proposed by the present invention can be converted into a program that is actually executed, and is stored in a non-transitory storage medium of a computer device (for example, a firmware, a flash drive, a solid state drive, an SD card, etc.). In this way, the high-definition television can execute the program corresponding to the image enhancement method of the present invention through the processor, thereby improving the dynamic range of the image, and the user can obtain better visual sensory satisfaction. It should also be mentioned that for the display devices that are ubiquitous today, such as computer screens, mobile phone screens, etc., the image enhancement method of the present invention can also be applied to enhance image details.

In summary, the present invention provides an image enhancement method that exhibits a good image dynamic range on a display of various specifications. Step by removing V channel image, using bidirectional filter, and limiting contrast adaptive histogram The image details of the color image in the bright and dark parts are clearly presented, which solves the problem that the dynamic range of the image on the low-contrast display is too low, and can meet the demand for high dynamic range of the high-definition television in the future.

While the invention has been described above by way of example, the invention is not intended to be limited thereby, the scope of the invention is defined by the scope of the appended claims.

S101~S106‧‧‧Steps

Claims (5)

An image enhancement method includes: acquiring an HSV color space of a color image; and extracting the color image corresponding to one H channel image, one S channel image, and one V channel image of the HSV color space; using a bidirectional filter The V channel image separates a base layer image and a detail image; the base layer image is processed by using a limited contrast adaptive histogram equalization step to obtain a contrast enhanced base layer image; the base layer image and the detail are contrast enhanced The image is combined to obtain a new V channel image; and the new V channel image is combined with the original H channel image and the S channel image to obtain a new color image after the color image is enhanced and contrasted. The image enhancement method of claim 1, wherein the detail image is obtained by subtracting the base image from the V channel image. The image enhancement method of claim 1, wherein the bidirectional filter is represented by the following relationship: ;as well as Where J _s is the result of processing a pixel s by a bidirectional filter, I _p and I _s are the intensity values of a pixel p and the pixel s, respectively, Ω is the entire image, and f and g are respectively in a spatial domain and A Gaussian smoothing function of an intensity domain, K _s is a normalization function. The image enhancement method of claim 1, wherein the limiting contrast adaptive histogram equalization step crops a histogram formed by pixels of the base image by a predetermined threshold to limit accumulation. The distribution function value (CDF) is magnified by the magnitude, and the value of the histogram that is cropped can be expressed by the following relationship: Where M x N is the total pixel value of the base layer image, L is the maximum image intensity value, α is a clipping factor, and S _Max is the maximum slope. The image enhancement method of claim 1, wherein the color image is presented on a display, and the contrast of the color image is higher than the contrast exhibited by the display.