JP2011254311A - Vehicle peripheral image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize color correction with a simple configuration which allows easy recognition of a preset object being present around a vehicle with higher precision as much as possible, even under such light source as poses strong color cast effect.SOLUTION: The vehicle peripheral image processor includes a light source estimating part 51 which estimates light source type as a light source from a white balance adjustment signal outputted from an onboard photographing unit, a color profile storing part 52 which stores color profiles generated for each light source type to have a characteristic in which such effect as a particular color in the photographed image is affected by a particular type of light source is reduced, and a color conversion part 54 which, if the inputted photographed image is the one photographed under the light source of particular type of light source, reads a color profile corresponding to the type of light source out of the color profile storing part 52, and corrects the color of the photographed image using the color profile.

Description

  The present invention relates to a vehicle periphery image processing apparatus that recognizes a specific object existing around a vehicle through a captured image.

  When the front of the vehicle is photographed with a camera mounted on the vehicle and a specific object (such as a stop position confirmation target or a road sign) is recognized from the captured image by image recognition, such object is There is a possibility of being illuminated by various types of light sources, and in particular, it becomes difficult to detect the contour of an object due to a color difference. In general, a technique called white balance adjustment is known to solve the problem of color cast caused by a light source. This white balance adjustment is one of image processing techniques for adjusting a captured image so that a display color recognized when a human directly looks at an object to be captured is also accurately reproduced in the captured image.

  For example, as an in-vehicle image recognition device employing such a white balance adjustment technology, a dominant light source outside the vehicle (based on a color temperature estimated from a photographed image of an achromatic member disposed at a predetermined position in the vehicle) A technique for estimating the color temperature of the main light source and reducing the influence of the light source on the captured image is known (see, for example, Patent Document 1). In other words, in order to estimate the color temperature of the light source outside the vehicle based on the captured image captured by the imaging unit mounted on the vehicle, an achromatic member is provided at a predetermined position in the vehicle. An image of the periphery of the vehicle is taken so as to include the sky. Estimate the color temperature of the image area of the achromatic member in the captured image, and determine the color temperature of one or more light sources existing in the image area of the sky in the captured image. presume. Then, based on each estimation result, the color temperature of the light source existing in the sky image area, which is within a predetermined color temperature determination range including the color temperature of the image area of the achromatic member, is determined as the final color temperature. It is assumed that In this technology, not only is it necessary to place an achromatic member at a predetermined position in the vehicle, but the achromatic member that becomes this color reference is contaminated or exposed to a local light source species. Therefore, it is impossible to accurately estimate the color temperature.

  There is known a video signal (photographed image) processing device that employs a white balance adjustment technique for performing good color reproduction correction without using an achromatic member attached to the vehicle body, regardless of the light source (for example, a patent) Reference 2). In this apparatus, white balance adjustment means for adjusting white balance of a video signal (captured image), color reproduction correction means for performing color reproduction correction on the video signal in accordance with color reproduction correction characteristics, and white balance adjustment A light source characteristic estimation unit that estimates a light source characteristic according to an adjustment state; a color reproduction correction characteristic control unit that variably controls a color reproduction correction characteristic in the color reproduction correction unit according to an estimation result of the light source characteristic estimation unit; Color reproduction correction characteristic storage means for storing a plurality of color reproduction correction characteristics respectively corresponding to different light source characteristics; and a plurality of color reproduction correction characteristics stored in the color reproduction correction characteristic storage means by the color reproduction correction characteristic control means. Among them, there is one that selects a color reproduction correction characteristic corresponding to the light source characteristic estimated by the light source characteristic estimation means. With this configuration, since the color reproduction characteristic corresponding to the estimated light source characteristic is selected from a plurality of color reproduction correction characteristics prepared in advance, the color reproduction of the entire captured image is improved regardless of the light source. However, since this apparatus intends to reproduce the color of the entire captured image based on the estimated light source characteristics, for example, when a strong color cast is generated by a very specific light source such as a sodium lamp light source, A sufficient effect cannot be obtained. As a result, even if a rough subject can be confirmed when a person sees the corrected photographed image, an object based on a color difference necessary for recognizing a stop position confirmation target or a road sign by image recognition. It is difficult to improve the accuracy of contour detection.

JP 2009-271112 A (paragraph number [0009-0021], FIG. 1) JP 2009-239323 A (paragraph number [0023-0046], FIG. 1)

  An object of the present invention is to provide a preset recognition target that exists in the vicinity of a vehicle with as high accuracy as possible even under a light source that exerts a strong color fogging effect such as a sodium lamp light source in view of the above-described prior art. It is an object of the present invention to provide a vehicle peripheral image processing apparatus that can realize color correction that facilitates object recognition with a simple configuration.

  In order to achieve the above object, a vehicle periphery image processing apparatus according to the present invention includes an imaging device and a white balance adjustment unit, and outputs a captured image of a vehicle periphery under a light source and a white balance adjustment signal for the captured image. A light source estimation unit for estimating a light source type indicating the type of the light source from the white balance adjustment signal, and the light source so as to have a characteristic that reduces an influence of a specific color in the photographed image by the light source type. A color profile storage unit that stores a color profile created for each type, and when the input captured image is a captured image under a light source indicated by the light source type, a color profile corresponding to the light source type is stored in the color profile. And color-correct the captured image using the color profile. And a color conversion unit.

  According to this configuration, the color influence that the specific color included in the vehicle peripheral image to be processed is affected by the light source is a color profile corresponding to the light source type estimated using the white balance adjustment signal that can be extracted from the photographing unit. Can be reduced by correcting the photographed image by the color conversion unit in which is set. For example, it is assumed that an important subject in the vehicle periphery image is a red lamp under a sodium lamp light source. In this case, it is necessary to correct the color of the red lamp so that it is as close to the original red as possible. In order to solve this problem, first, a color profile is created and stored so that the original color is restored as much as possible even when the red color changes under a sodium lamp light source. Then, the sodium lamp light source is estimated as a light source type from the white balance adjustment signal, and the above color profile for returning the red color affected by the discoloration under the sodium lamp light source to the original color as much as possible is read. Can be color corrected.

  If the color effect (influence on color) that a specific color on a specific recognition target object such as a road sign or a stop marker in the original photographed image is affected by the color correction is reduced by color correction, By using the captured image, the image recognition accuracy of such a recognition object is improved. Therefore, it is advantageous to provide the vehicle peripheral image processing apparatus according to the present invention with an image recognition module that recognizes a recognition object set in advance in the captured image output from the color conversion unit.

  When a road sign, a stop marker, or the like is recognized as an image, its contour detection or position detection is often performed with a binary image. Therefore, when performing color correction such as color fog removal, it is not preferable to lower the dynamic range by lowering the luminance level of the original photographed image. Therefore, in one preferred embodiment of the present invention, the color profile of the light source type maintains the luminance level by reducing the specific color affecting the recognition target and increasing the complementary color of the specific color. It is created to have the characteristics to

  Since the color influence such as color cast by the light source depends on the color illuminated by the light source, if there are a plurality of characteristic colors that characterize the recognition object, a color profile is required for each characteristic color. Accordingly, in one preferred embodiment of the present invention, the color conversion unit reads a color profile corresponding to the light source type and a characteristic color characterizing the recognition target object from the color profile storage unit, and The captured image is color-corrected using a color profile.

  An object that is effectively recognized by the present invention is a marker composed of, for example, a combination of different colors (for example, two colors) that is used for inspection of mounting accuracy of a vehicle-mounted camera or the like and target setting of a parking position. Since colors (including approximate colors) to be color-corrected are limited, effective color influence can be removed. Further, among the light sources that illuminate the recognition object around the vehicle, a sodium lamp can be cited as a particularly intense color cast. For this reason, it is preferable to have a color profile that uses a sodium lamp as a light source species at a minimum.

It is explanatory drawing of the first half of the color correction principle which correct | amends the color influence by the light source employ | adopted with the vehicle periphery image processing apparatus by this invention. It is explanatory drawing of the second half of the color correction principle which correct | amends the color influence by the light source employ | adopted with the vehicle periphery image processing apparatus by this invention. The functional block diagram of the parking assistance apparatus which employ | adopted the vehicle periphery image processing apparatus by this invention is shown. It is a schematic diagram explaining the picked-up image containing the parking lot line and parking marker used for parking assistance. It is a flowchart which shows the flow of control of parking assistance. It is a flowchart which shows the flow of the color correction by a color correction module.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, the principle of the color correction technique that is used in the vehicle peripheral image recognition apparatus according to the present invention and corrects the color influence (effect related to color) caused by the light source will be described with reference to FIGS.

Here, the recognition object 2 is assumed to be an object characterized by two different colors in order to simplify the description. As shown in FIG. 1A, one of the two colors (original colors including approximate colors) is named as the first color and expressed in C1 (R1, G1, B1) (RGB color system). Yes. The other of the two colors is named the second color and is represented by C2 (R2, G2, B2). Note that the term “original color” represents an original color under ideal sunlight, but here, such a color is not limited to a narrow range, but a predetermined color including an approximate color of the original color is also included. It is used as a phrase having a color range of. Further, here, the characteristic colors characterizing the recognition object 2 are the first color and the second color, and accurate position detection is possible through detection of the boundary line between the first color and the second color. . When the recognition object 2 is illuminated by a light source such as a tungsten lamp or a sodium lamp, the first color and the second color change to different colors (change colors including approximate colors). The color in which the first color is changed is named the third color and is represented by C3 (R3, G3, B3). The color in which the second color is changed is named the fourth color and is represented by C4 (R4, G4, B4). Each color change can be expressed as follows using a relational expression expressing the color change (transition) in a matrix.
C1 (R1, G1, B1) ・ M1 => C3 (R3, G3, B3) (1)
C2 (R2, G2, B2) ・ M2 => C4 (R4, G4, B4) (2)
A schematic diagram visualizing such a color change on the color system is shown in FIG. That is, the coordinate position of the original color is moved to the coordinate position of the change color by illumination with the light source, and the direction and distance of the movement are represented by conversion matrices: M1 and M2.
Next, let us consider color correction of the recognition target object 2 that is affected by the color of the light source so that it becomes the original color on the captured image. This correction can be realized by obtaining transformation matrices: M1 and M2 inverse matrices: M1 ⁻¹ and M2 ⁻¹ . In other words, the movement from the first color to the third color on the color system and the movement from the second color to the fourth color on the color system due to the color effect of the light source becomes the original color by reversing the movement in reverse. . This can be shown as follows.
C3 (R3, G3, B3) ・ M1 ^-1 => C1 (R1, G1, B1) (3)
C4 (R4, G4, B4) ・ M2 ^-1 => C2 (R2, G2, B2) (4)
A schematic diagram visualizing such color correction on the color system is shown in FIG.
A schematic diagram visualizing such color correction on the color system is shown in FIG.
In general, since it is difficult to obtain an accurate inverse matrix, it may be obtained approximately or experimentally or both. The color profile (color matrix) created so as to correspond to M1 ^-1 and M2 ^-1 obtained in this way has characteristics that at least reduce the color influence of the color target 2 on a specific light source. It will be.
In addition, a common transformation matrix is preferably created by optimally approximating a plurality of transformation matrices created individually for each color (first color and second color), and both colors (first and second colors) are obtained from the inverse matrix of the common transformation matrix. A color profile that matches one color and the second color may be created. In addition, although substantially the same, a plurality of individually created inverse matrices are preferably optimally fused into one inverse matrix, and both colors (first color and second color are selected from this fused inverse matrix. ) May be created.

When a plurality of light sources that have a color effect such as strong color cast have to be considered, it is necessary to create a color profile as described above for each light source and store it in a selectable manner. Furthermore, it is necessary to perform light source estimation as to what kind of light source is illuminated from the captured image to be processed. Although it is known that this light source estimation is performed from various image feature amounts of a photographed image to be processed, statistical calculation values such as luminance histograms and color component histograms, various image feature amounts are known. The calculation for obtaining is accompanied by a considerable burden. For this reason, the present invention uses a white balance adjustment function attached to a photographing unit including a digital camera as a component, and uses a white balance adjustment signal from the photographing unit as an input signal to calculate a light source type (estimation). ) Method. That is, as schematically shown in FIG. 2, the adjustment signal for white balance adjustment executed at the time of shooting of the shot image to be processed is digitized, and a preliminary calculation is performed as necessary, so that the light source estimation algorithm is processed. Given as input data. Typical input data includes a luminance average value of a target captured image, R (red): average value, G (green): average value, B (blue): average value, and the like. When such input data based on the white balance adjustment signal is given, a light source estimation algorithm calculation is activated, and a light source type is calculated (estimated) and output. As this calculation algorithm, a rule-based mechanism (for example, calculation such as comparing input data with a predetermined threshold value) may be adopted, but a weighting factor is set by learning in advance. Thus, a mechanism like a neural network for producing an inference operation unit is convenient.
In any case, when such an inference calculation unit is constructed, the light source type is output only by inputting the white balance adjustment signal or data based on the white balance adjustment signal. Next, a color profile (inverse matrix) adapted to the output light source type is selected and output from the color profile storage unit, and the captured image is color-corrected using the output color profile, whereby the recognition object 2 is detected. A captured image that facilitates identification and recognition of the characteristic color being characterized is obtained.

  Next, one embodiment of a vehicle peripheral image processing apparatus according to the present invention, which employs the principle of the color correction technique for correcting the color influence due to the light source described above, will be described with reference to the drawings. In this embodiment, the vehicle periphery image processing device is incorporated as an image processing device in a parking assistance system. FIG. 3 is a functional block diagram schematically showing the configuration of such a parking support system. The parking assistance system includes a vehicle surrounding image processing device 3, a parking assistance ECU 7, and a man / machine interface unit 8, and the vehicle surrounding image processing device 3 includes a photographing unit 4, a color correction module 5, and an image recognition module 6. The man-machine interface unit 8 is connected to a monitor, a touch panel, a speaker, and various sensors.

  The imaging unit 4 photoelectrically converts the image of the subject imaged by the imaging lens 41 and outputs an image signal, and the image signal output from the imaging element 3 is converted into three primary colors of R, G, and B. Alternatively, a color separation unit 43 that separates the color difference signals and a white balance adjustment unit 44 that performs white balance adjustment on the color-separated signals are provided. Since the RGB signal output from the color separation unit 43 has a white or gray color due to the spectral characteristics of the light source, the white balance is adjusted by the white balance adjustment unit 44. Since white balance adjustment is well known, a detailed description is omitted here. For example, the average value of the entire photographed image for each of R, G, and B, and the weight value based on the calculated value obtained from the average value by a predetermined calculation formula, etc. A balance adjustment value is obtained, and the pixel value of the captured image is corrected using the white balance adjustment value.

The color correction module 5 includes a recognized feature color setting unit 50, a light source estimation unit 51, a color profile storage unit 52, a color profile selection unit 53, and a color conversion unit 54. The recognition feature color setting unit 50 sets a feature color that characterizes the recognition target in the captured image that is the processing target. The color correction module 5 performs correction such that the characteristic color set here is preferentially favorably corrected, and performs average color correction when such characteristic color is not set. The light source estimation unit 51 inputs data based on the white balance adjustment signal sent from the white balance adjustment unit 54, and outputs the estimated light source type through operations such as weight calculation and rule calculation. The weight calculation here is, for example, a calculation that assigns a weight coefficient to each input parameter and corrects the weight coefficient by iterative learning to minimize the error in the output result, as used in neural networks, for example. It is a general term for operations using expressions. A rule operation is a general term for operations that derive a result based on a predetermined rule such as an ifthen statement. Examples of the content of the rule itself include “if the B color component value is higher than a predetermined value and the R and G component values are lower than the predetermined value, the possibility of an orange lamp is higher than a predetermined percentage”.
In addition, although a tungsten lamp, a sodium lamp, a fluorescent lamp, an LED lamp, etc. are mentioned as a light source type (type of light source), a combination with each other may be used as the light source type estimated here. Also, the same light source type may be classified according to the intensity of each lamp and the magnitude of color influence.

  The color profile storage unit 52 stores a color profile used for the color correction described with reference to FIG. This color profile is a correction table for reducing as much as possible the color influence such as the color cast on the characteristic color characterizing the recognition object in the photographed image. In this embodiment, each color profile is created for each combination of a specific light source type and a specific color configuration of the recognition object. Further, the color profile is stored in a database so that the light source type and the characteristic color of the recognition target object can be searched and extracted as search keywords. The color profile selection unit 53 selects a suitable color profile from the light source type estimated by the light source estimation unit 51 and the preset characteristic color of the recognition target object, and gives the color profile to the color conversion unit 54. The color conversion unit 54 performs color correction of the captured image using the color profile selected by the color profile selection unit 43.

  The image recognition module 6 has a function of recognizing a preset recognition target object in the color-corrected captured image output from the color correction module. In this embodiment, the recognition target object is generally white (characteristic color) parking partition lines L1 and L2 bounding the parking area as illustrated in FIG. 4, and two specific colors (characteristic color). ) Is a parking marker PM characterized by Therefore, the image recognition module 6 includes a parking lot line recognition unit 61 that detects the parking lot lines L1 and L2, and a parking marker recognition unit 62 that detects the parking marker PM. Since the image recognition algorithm for recognizing the recognition object as described above from the captured image is well known, description thereof is omitted here, but the white color characterizing the parking lot lines L1 and L2 and the two colors characterizing the parking marker PM ( For example, red and blue) are set as recognition feature colors by the recognition feature color setting unit 50.

  The parking assist ECU 7 includes a target parking position setting unit 71, a vehicle position calculation unit 72, a parking route calculation unit 73, and a guidance unit 74, as is well known per se. The target parking position setting unit 71 sets the target parking position based on the recognition information regarding the parking lane lines L1 and L2 and the parking marker PM recognized by the image recognition module 6. The vehicle position calculation unit 72 detects the position of the vehicle based on sensor information obtained by a distance sensor, a speed sensor, a steering angle sensor, and the like in addition to the recognition information. The parking route calculation unit 73 calculates a parking route to the target parking position based on the target parking position and the vehicle position. The guide unit 74 controls the accelerator ECU, the steering ECU, the brake ECU, and the like according to the parking route to guide the vehicle to the target parking position.

  The operation of the parking support apparatus when the vehicle 10 equipped with the parking support system configured as described above parks in the parking section E as shown in FIG. 4 will be described with reference to the flowchart of FIG.

  In this state, a place where the driver stops the vehicle to park in the parking section E is set as a parking start position (# 01). A photographed image (here, a vehicle rear peripheral image) is acquired by the photographing unit 4 of the vehicle stopped at a predetermined position. (# 02). When the parking start position is the position illustrated in FIG. 4, the captured image is as illustrated in FIG. This captured image is also displayed on a monitor provided in the passenger compartment. In the upper left of this captured image, a parking section E and a parking marker PM partitioned by the parking section line L1 and the parking section line L2 are included. The obtained captured image is subjected to the color correction process shown in FIG. 6 (# 03).

  In the color correction process, the captured image (RGB color image data) input to the color correction module 5 is subjected to basic image processing such as correction of image distortion caused by lens characteristics (# 31). Next, in order to estimate under what kind of light source the photographed image was photographed, it is now input based on the white balance adjustment signal sent from the white balance adjustment unit 44 of the photographing unit 4 Input data for estimating the light source in the captured image is created (# 32). A light source type estimation calculation is performed using the created input data as an input value to output a light source type (# 33). The color configuration of the recognition target object to be processed is read (# 34). For example, when the recognition target is the parking section E, the characteristic colors are gray of the concrete ground and white of the white line, so that the color configuration is gray and white. Further, when the recognition target object is the parking marker PM, the characteristic colors are red and blue, and the color configuration is red and blue. A matching color profile is selected using the output light source type and the read color configuration as a search keyword (# 35). Using the selected color profile, color correction is performed on the captured image (# 36). The color-corrected captured image is developed in the working memory and used for the following image recognition / parking support process.

In the image recognition / parking support process, first, the parking lot lines L1 and L2 are detected (# 04). Here, the detection of the parking lot line is performed according to the following procedure.
(1) A viewpoint conversion process is performed by converting a captured image from world coordinates to image coordinates.
(2) For an image subjected to viewpoint conversion processing, for example, a 3 × 3 spatial filter is scanned, and differential processing is performed on each spatial filter to obtain luminance difference information. Based on the acquired luminance difference information, a parking lot line candidate is extracted from the luminance difference exceeding the threshold value for each spatial filter.
(3) When a parking lot line candidate is extracted, a parking lot line is detected from the parking lot line candidate using Hough conversion or the like. When the parking lot line is extracted, the coordinate values of the ends of the parking lot lines L1 and L2 are acquired based on the detected parking lot line.

  When the parking lot line detection process ends, a target parking position is set (# 05). The target parking position is set so that the vehicle is parked at the center of the parking section based on, for example, data on the parking section lines L1 and S2. When the target parking position is set, a parking route for parking is calculated based on information on the current vehicle position and information on the target parking position (# 06), and guidance of the vehicle 10 to the target parking position is performed. This is executed (# 07). When the guidance starts, a new captured image is input (# 08), color correction processing is performed (# 09), and the position detection of the parking marker PM which functions as a parking stop marker here from the color-corrected captured image. Is performed (# 10). The position detection of the parking marker PM can also be performed by a method similar to the parking lot line. When the position detection of the parking marker PM is completed, it is checked whether the vehicle has reached the parking stop position based on the positional relationship between this position and the vehicle (# 11). If the vehicle has not reached the parking stop position (# 11 No branch), the process returns to step # 07 to continue the guidance. If the vehicle has reached the parking stop position (# 11 Yes branch), the vehicle is stopped and the parking support is terminated (# 12).

  There are various recognition objects to be recognized using the captured images around the vehicle acquired by the imaging unit 4 which is an in-vehicle camera, but the accuracy is improved by using the color correction according to the present invention. In particular, there are lane marks such as traveling lanes and passing lanes as recognition objects that are required to be accurately recognized. The road on which the lane mark is drawn is exposed to various light sources depending on the type, for example, an expressway, a tunnel road, a general road, and the like, so that the application of the present invention is effective. Even when the lane mark is a recognition target, the color correction process as described above is performed on the acquired captured image, so that the lane mark for each color is not greatly affected by the type of light source illuminating the lane mark. Can be recognized. In such an embodiment, the configuration of the color correction module 5 is substantially the same, and the image recognition module 6 includes, for example, a yellow lane mark recognition function and a white lane mark recognition function. The image recognition module 6 first converts the color-corrected RGB image sent from the color correction module 5 into color system image data such as CIELab or HSV that allows easy color comparison. Here, it is assumed that the HSV color system is adopted. In order to recognize the existence range of the yellow line (passing lane) from the converted image data, the threshold value of the yellow saturation value that divides the road area and the yellow line area is calculated, and the yellow line is calculated using this threshold value. Are extracted, and a binary image of a yellow line is generated. From the generated binary image of the yellow line, the position of the overtaking lane is recognized through the contour detection process as described above. In order to recognize the existence range of the white line (running lane), the threshold value of the white luminance value that divides the road area and the white line area is calculated, and the white line is extracted using this threshold value. The binary image is generated. From the generated binary image of the white line, the position of the traveling lane is recognized through the contour detection process as described above.

  As described above, when recognizing an image using the acquired photographed image, the light source type affecting the photographed image is estimated as a pre-process, and the color influence due to the estimated light source type is particularly recognized. Reduced color correction is performed with respect to the color characterizing the object. Thereby, even under the situation where the influence of the light source fluctuates with the movement of the vehicle, appropriate color correction can be performed, and the accuracy of the image recognition can be improved.

[Another embodiment]
(1) Each of the functional units in the above-described vehicular image processing apparatus, particularly the color correction module 5, indicates the sharing as a function, and is not necessarily provided independently, and any combination thereof is combined. One functional unit may be provided, or each functional unit described above may be further divided.
(2) In the above-described embodiment, the recognition objects of the image recognition module 6 are the parking lot lines L1 and L2 and the parking marker PM in the parking lot, and white lines and yellow lines drawn on the road. . However, the recognition target handled in the present invention may be a target for stopping at the battery charging station, another road sign, or a target related to vehicle travel. At this time, the characteristic color that characterizes the recognition object may be set automatically or manually in the recognition characteristic color setting unit 50 according to the recognition object to be recognized.
(3) A sodium lamp light source is a light source type that has the strongest influence of color cast in a photographed image, and is often used for lighting in tunnels and parking lots. Therefore, a color profile limited to the sodium lamp light source and other general white balance adjustment color profiles may be created to simplify the profile configuration.
(4) Since a light source that causes strong color cast has color characteristics that are biased toward a specific color component, when creating a color profile for such a light source type, the specific color is reduced and a complementary color of the specific color is applied. Increasing the luminance level by increasing the value is advantageous in the subsequent binarized image processing.

  The present invention can be widely used for an image processing technique for reducing the color influence of a light source exerted on a subject characterized by a combination of different colors.

2: recognition object 4: photographing unit 42: photographing element 43: color separation unit 44: white balance adjustment unit 5: color correction module 50: recognition characteristic color setting unit 51: light source estimation unit 52: color profile storage unit 53: color Profile selection unit 54: color conversion unit 6: image recognition module 61: parking lot line recognition unit 62: parking marker recognition unit 7: parking assist ECU
8: Man-machine interface units L1, L2: Parking lot line PM: Parking marker

Claims (6)

An in-vehicle imaging unit that includes an imaging element and a white balance adjustment unit, and outputs a captured image of the vehicle periphery under a light source and a white balance adjustment signal for the captured image;
A light source estimation unit that estimates a light source type indicating the type of the light source from the white balance adjustment signal;
A color profile storage unit that stores a color profile created for each light source type so as to have a characteristic that reduces the influence of a specific color in the photographed image by the light source of the light source type;
When the input photographed image is a photographed image under the light source of the light source type, color conversion corresponding to the light source type is read from the color profile storage unit, and the photographed image is color-corrected using the color profile. And
A vehicle periphery image processing apparatus.   The vehicle periphery image processing apparatus according to claim 1, further comprising an image recognition module that recognizes a preset recognition target object in the captured image output from the color conversion unit.   The vehicle periphery image processing device according to claim 2, wherein the color profile of the light source type has a characteristic of maintaining a luminance level by reducing a specific color that affects a recognition target object and increasing a complementary color of the specific color.   The color conversion unit reads a color profile corresponding to the light source type and a characteristic color that characterizes the recognition target object from the color profile storage unit, and color-corrects the captured image using the color profile. The vehicle periphery image processing device according to 3.   5. The marker according to claim 2, wherein the recognition object is a marker composed of a combination of different colors, and the image recognition module determines the position of the marker in the captured image by identifying and determining different color regions of the marker. The vehicle periphery image processing device according to one item.   The vehicle periphery image processing device according to any one of claims 1 to 5, wherein the light source type is a sodium lamp light source.

Images