JP4523945B2 - Image processing method, image processing apparatus, and image processing program - Google Patents

Description

  The present invention relates to a technique for detecting a change in a natural environment by processing a captured image.

  Conventionally, observation of weather information and observation of changes in the natural environment have been performed by installing fixed sensors. For example, a change in the natural environment is detected by detecting a change in color due to a secular change in a building, a signboard, a sign, or the like in the natural environment. Specifically, changes in the natural environment are detected by examining the colors of houses, buildings walls, roads, bridges, etc. that have changed due to rust, corrosion, chemical changes, and the like.

It is also important to manage the visibility of signs, etc. to a certain level or more, as it is necessary to repair or replace signs, traffic signs, etc. due to weathering due to rust, corrosion, chemical changes, etc. It is. In addition, there are the following as technical documents related to the present application.
“Color Conversion Formulas”, [online], System Planning Laboratory, [Search December 15, 2006], Internet <URL: http: // image-d. isp. jp / commentary / color_formula />

  However, since it may be difficult for an amateur to immediately determine the degree of weathering, it is necessary to perform quantification processing of an image obtained by photographing a survey target.

  On the other hand, with the recent improvement in performance of mobile terminals such as camera-equipped mobile phones, it has become possible to acquire information in various places by using mobile terminals in areas where it is difficult to install fixed sensors.

  However, when photographing outdoors, it is difficult to obtain an image with a correct color due to various lights, noises, and shadows. White balance is most widely used as an image correction method. However, when white balance is used, there is a problem that it becomes ineffective when the center of shooting is shifted even a little in a dark place or a place where light is partially irradiated.

  Further, γ correction is known as a means for correcting brightness. The γ correction is performed when the photographer artificially operates the setting menu of the camera and adjusts the image to have the optimum brightness.

  This invention is made | formed in view of the above, The place made into the subject exists in detecting easily the grade of the weathering of test object.

  Another object of the present invention is to easily obtain correct color information even outdoors.

An image processing method according to the first aspect of the present invention includes a step of inputting RGB image data obtained by photographing an inspection object by an input unit, a step of correcting brightness of the RGB image data by γ correction by a correction unit, The RGB image data whose brightness has been corrected by the conversion means is converted into an HSV color space model, and the hue information is extracted, and the hue information and information corresponding to the hue information are stored in association with each other by the matching means. The hue information matching the hue information extracted in the step of extracting the hue information is retrieved from the storage means, the information corresponding to the hue information is read from the storage means, and the information corresponding to the hue information is output by the output means. A test chart composed of a plurality of parts having the same area and different colors in a γ correction value used for γ correction. For RGB image data captured under the same conditions as were taken inspection object, and step a, brightness corrected the RGB image data to HSV color space model to correct the brightness of the RGB image data by γ correction The steps of converting the image into a hue and extracting the hue distribution and the step of determining the number of hue peaks of the extracted hue distribution and the frequency of pixels at each peak are repeatedly executed while gradually increasing the γ correction value from a predetermined initial value. Then, the number of hue peaks in the extracted hue distribution is the same as the number of colors in the test chart, and the calculation is performed by obtaining a γ correction value that makes the frequency number of pixels in each peak equal.

  In the present invention, the brightness of the input RGB image data is corrected, converted to another color space model to extract the hue information, and the hue information and information indicating the degree of weathering corresponding to the hue information By searching for hue information matching the extracted hue information from the storage means stored in association with each other, it is possible to easily know the degree of weathering of the object copied in the input RGB image data. .

  In the present invention, when a test chart is input as RGB image data and the hue information extracted by correcting the brightness with a predetermined correction value is not the desired hue information, the correction value is changed and the brightness is changed again. If the desired hue information is obtained as a result of correcting the brightness and correcting the brightness, the optimum correction value is used as the correction value at that time, so that the optimum value can be easily obtained without bothering the user. A correction value can be calculated, and correct hue information can be obtained.

An image processing apparatus according to a second aspect of the present invention includes a storage unit that associates and stores hue information and information corresponding to the hue information, an input unit that inputs RGB image data obtained by photographing an inspection object, and an RGB image Correction means for correcting the brightness of the data by γ correction, conversion means for converting the RGB image data corrected by the correction means into an HSV color space model, and extracting hue information; hue information extracted by the conversion means; A matching unit that matches the hue information stored in the storage unit and reads information corresponding to the hue information from the storage unit; and an output unit that outputs information corresponding to the hue information. RGB image data of the correction value, taken at the same conditions as the test chart including a plurality of sites of different colors in the same area that is input from the input means and photographed test object Then, the brightness is corrected by γ correction by the correction means, and the hue distribution is extracted by converting to the HSV color space model by the conversion means, while repeatedly increasing the γ correction value from a predetermined initial value, The calculation is performed by obtaining a γ correction value in which the number of hue peaks of the extracted hue distribution is the same as the number of colors in the test chart and the frequency number of pixels at each peak is equal.

  An image processing program according to a third aspect of the present invention causes a computer to execute each step in the image processing method.

  According to the present invention, it is possible to easily detect the degree of weathering of an inspection target. Also, correct color information can be easily obtained even outdoors.

  Hereinafter, embodiments 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 the present embodiment. The image processing apparatus shown in FIG. 1 includes an image input unit 11, a correction unit 12, a conversion unit 13, a collation unit 14, an output unit 15, and a storage device 16. The image processing apparatus 1 can be configured by a computer including an arithmetic processing device, a storage device, a memory, and the like, and processing of each unit is executed by a program. This program is stored in the storage device 16, and can be recorded on a recording medium or provided through a network. In addition, by configuring the image processing apparatus 1 with a camera-equipped mobile terminal or the like, it is possible to acquire information at various places.

  The image processing apparatus 1 is characterized in that an RGB image obtained by photographing an object by the image input unit 11 is input, hue information is extracted by converting the RGB image into an HSV model, and the object's object is extracted. The storage device 16 that stores data that associates the hue information with the information representing the state of the object corresponding to the hue information is obtained by matching the data similar to the extracted hue information to obtain the state of the object. Knowing it, it is possible to detect changes in the natural environment by judging the state of the object. Hereinafter, processing in each unit will be described in detail.

  The image input unit 11 inputs RGB image data obtained by capturing an object with a camera or the like, and stores it in the storage device 16. Since most camera-equipped mobile terminals shoot images with R (red), G (green), and B (blue) light receiving elements, the obtained image data is RGB image data in which each pixel is represented by RGB elements. It becomes.

The correction unit 12 reads the RGB image data from the storage device 16, performs γ correction, and stores it in the storage device 16. Since the photographed image tends to be an image affected by shadows or gloss, it is appropriate to adjust the brightness and obtain an image that is photographed with uniform ambient light. The brightness is adjusted by converting the RGB value for each pixel with an exponential function. In an image sensor such as a CCD camera, the input light amount E and the output value D have an exponential function relationship called a γ characteristic shown in the following equation (1).

The γ correction is to correct this γ characteristic. Specifically, the γ correction is performed by using the following conversion formula for the RGB value for each pixel.

  The pixel value takes a value from 0 to 255. It becomes bright when γ> 1, and dark when γ <1. Thus, by applying γ correction to the input RGB image data, it is possible to suppress the influence of sunlight, shadows, etc., and obtain the original color information of the object. An appropriate γ correction value calculation method will be described later.

  The conversion unit 13 reads out RGB image data from the storage device 16, converts it into an HSV color space model, and extracts hue information. As shown in FIG. 2, the RGB color space model is an orthogonal coordinate system in which R, G, and B are assigned to each axis. On the other hand, the HSV color space model is a color space including three components of hue (Hue), saturation (Saturation), and lightness (Value), and can be expressed in a conical coordinate system as shown in FIG. Hue represents a hue such as reddish, yellowish, and blued. Saturation represents a color intensity, vividness, and lightness represents the brightness of the color.

  The matching unit 14 searches the similar information by matching the hue information extracted by the conversion unit 13 with the hue information stored in the storage device 16. The storage device 16 stores the hue information of the object in association with the information indicating the state of the object corresponding to the hue information, and searches the storage device 16 for hue information similar to the extracted hue information. By doing so, information representing the state of the photographed object can be obtained. The information representing the state of the object is information such as information representing the degree of weathering of the object and information representing the state of progress of rust of the object, for example. Specifically, the hue distribution is used as the hue information, and the storage device 16 associates the degree of rust with information indicating the progress of rust, such as none, weak, or strong, according to the redness ratio of the hue. Store it. As examples of the hue information of the target object and the information indicating the state of the target object, information that associates the degree of brown with the turbidity of water, information that associates the sky color with the degree of air pollution, and the like can be considered.

  The output unit 15 reads out information representing the state of the object corresponding to the hue information matched by the matching unit 14 from the storage device 16 and outputs the information.

  Next, the flow of processing of the image processing apparatus 1 will be described using the flowchart shown in FIG. The image input unit 11 inputs RGB image data obtained by photographing the inspection object (step S41), and the correction unit 12 performs γ correction on the photographed RGB image data so that the brightness of the RGB image data is set to an appropriate brightness. It corrects (step S42). The conversion unit 13 converts the corrected RGB image data into an HSV color space model (step S43), and extracts hue information (step S44).

  Subsequently, data similar to the hue information extracted in step S44 is obtained by matching from the storage device 16 that stores data in which the hue information of the object and information indicating the state of the object corresponding to the hue information are associated. (Step S45), the output unit 15 outputs information indicating the state of the object corresponding to the input image (step S46). By referring to the output information, it is possible to know the degree of weathering of the object and to detect changes in the natural environment.

  Next, an appropriate correction value calculation method used in the correction unit 12 will be described.

  FIG. 5 is a diagram showing a test chart used for calculating an appropriate correction value. The test chart 500 shown in the figure is composed of four color portions denoted by reference numerals 501, 502, 503 and 504, and the area occupied by each portion is the same. When the test chart 500 is photographed outdoors and the hue information is extracted, the original hue distribution cannot be obtained due to the effects of various noises such as sunlight and shadows, and the hue and the frequency of pixels with respect to it vary. The graph is as shown in FIG. In the graph shown in FIG. 6, the horizontal axis represents hue, and the vertical axis represents pixel frequency. The hue can be set, for example, in 360 levels from 0 to 359.

  Since the test chart 500 is composed of four color parts having the same area, the hue distribution of the original test chart 500 is four as shown in FIG. The frequency number is also the same.

  In the image processing apparatus 1, an appropriate correction value is calculated by inputting the test chart 500 and changing the γ correction value so that the extracted hue information becomes the original hue distribution of the input test chart 500. Hereinafter, calculation of the correction value will be described.

  FIG. 8 is a flowchart showing a flow of processing for obtaining an appropriate correction value. First, a desired number of peaks are input using an input device (not shown) such as a keyboard or a mouse (step S81). For example, since the test chart 500 is composed of four colors, the peak number is input as four. The test chart 500 is photographed, and the photographed RGB image data is input by the image input unit 11 and stored in the storage device 16 (step S82).

  Next, the correction unit 12 reads the RGB image data of the test chart 500 taken from the storage device 16, and performs γ correction with γ = 0.0 as an initial value (step S83). The conversion unit 13 converts the RGB image data subjected to γ correction into an HSV color space model (step S84), and extracts a hue distribution (step S85).

  It is determined whether the number of hue peaks of the extracted hue distribution is the same as the input peak number, and the frequency number of pixels in each peak satisfies a predetermined condition (step S86). For example, since the area of each color is the same in the test chart 500, a desired hue distribution is obtained when the frequency numbers of pixels at each peak are substantially equal.

  When the desired hue distribution is obtained, the γ correction value at that time is stored as an appropriate γ correction value in the memory or the storage device 16 provided in the correction unit 12 (step S87). If the desired hue distribution is not obtained, the correction value is changed, and the process proceeds to step S83 again. For example, the γ correction value is increased by 0.01 and the γ correction is performed again.

  As described above, by photographing and inputting the test chart 500 and calculating the γ correction value that can obtain the original hue distribution of the test chart 500, it is possible to suppress variations in outdoor brightness conditions. Further, by using the test chart 500, an appropriate γ correction value can be easily calculated without bothering the user.

  Therefore, according to the present embodiment, the correction unit 12 adjusts the brightness by γ-correcting the RGB image obtained by photographing the object, and the conversion unit 13 converts the image into an HSV color space model to extract the hue information. Then, the matching unit 14 searches the similar hue information by matching the extracted hue information with the hue information stored in the storage device 16, and indicates information indicating the degree of weathering of the object corresponding to the hue information. By reading out, it becomes possible to know the degree of weathering simply by photographing the object.

  According to the present embodiment, the optimal γ correction value is easily calculated by photographing the test chart 500 and changing the γ correction value and correcting the brightness until desired hue information is obtained. Therefore, correct hue information can be obtained without being influenced by the environment.

It is a block diagram which shows the structure of the image processing apparatus in one embodiment. It is explanatory drawing which shows RGB color space. It is explanatory drawing which shows HSV color space. It is a flowchart which shows the process of the image processing apparatus shown in FIG. It is a figure which shows the test chart used in order to calculate (gamma) correction value. It is a graph which shows hue distribution when the test chart shown in FIG. 5 is imaged. 6 is a graph showing an original hue distribution of the test chart shown in FIG. 5. 2 is a flowchart illustrating a process of calculating a γ correction value by the image processing apparatus illustrated in FIG. 1.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Image processing apparatus 11 ... Image input part 12 ... Correction | amendment part 13 ... Conversion part 14 ... Collation part 15 ... Output part 16 ... Memory | storage device 500 ... Test chart

Claims (3)

A step of inputting RGB image data obtained by photographing an inspection object by an input means;
Correcting the brightness of the RGB image data by γ correction by a correction means;
Converting the RGB image data whose brightness has been corrected by a conversion means into an HSV color space model, and extracting hue information;
The matching means searches the hue information that matches the hue information extracted in the step of extracting the hue information from the storage means that stores the hue information and the information corresponding to the hue information in association with each other, and stores the hue information in the hue information. Reading corresponding information from the storage means;
Outputting information corresponding to the hue information by output means;
Have
The γ correction value used for the γ correction is
Correcting the brightness of the RGB image data by γ correction for RGB image data obtained by photographing a test chart consisting of a plurality of parts of the same area and different colors under the same conditions as the subject to be examined ; Converting the RGB image data of which the color correction has been performed into an HSV color space model and extracting a hue distribution; and determining a hue peak number of the extracted hue distribution and a pixel frequency number at each peak. The γ correction value is such that the number of hue peaks in the extracted hue distribution is the same as the number of colors in the test chart, and the frequency number of pixels at each peak is equal, repeatedly executed while gradually increasing the γ correction value from a predetermined initial value. An image processing method characterized by calculating by calculating. Storage means for associating and storing hue information and information corresponding to the hue information;
Input means for inputting RGB image data obtained by photographing an inspection object ;
Correction means for correcting the brightness of the RGB image data by γ correction;
Conversion means for converting the RGB image data corrected by the correction means into an HSV color space model and extracting hue information;
A matching unit that matches the hue information extracted by the conversion unit and the hue information stored in the storage unit, and reads information corresponding to the hue information from the storage unit;
Output means for outputting information corresponding to the hue information;
Have
The γ correction value used for the γ correction is
With respect to RGB image data obtained by photographing a test chart composed of a plurality of parts of different colors with the same area and inputted from the input means under the same conditions as those obtained by photographing the inspection object , brightness is corrected by γ correction by the correction means. Correction, conversion to the HSV color space model by the conversion means and extraction of the hue distribution are repeatedly performed while gradually increasing the γ correction value from a predetermined initial value, and the number of hue peaks of the extracted hue distribution is calculated. An image processing apparatus, wherein the number is calculated by obtaining a γ correction value that is the same as the number of colors in the test chart and that equals the frequency number of pixels at each peak.   An image processing program for causing a computer to execute each step in the image processing method according to claim 1.

Images