JP2007193702A - Image processing device and image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing device and an image processing method that can accurately determine the presence of vehicle lights by discriminating lights of vehicles traveling at night from road surface reflections. <P>SOLUTION: An image processing part 24 binarizes a captured image captured by a video camera 1, and connects adjacent areas in the binary image by labeling to extract vehicle light candidate areas. The image processing part 24 calculates the roundness of each extracted vehicle light candidate area, determines whether or not the calculated roundness is at least a threshold, and identifies the vehicle light candidate area as a vehicle light area if the roundness is at least the threshold. If the roundness is smaller than the threshold, the image processing part 24 identifies the vehicle light candidate area as a road surface reflection and removes it from the captured image. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an image processing apparatus that performs processing for determining the presence of a vehicle light based on a captured image obtained by imaging an area including a road, and particularly distinguishes a vehicle light from night road surface reflection. The present invention relates to an image processing apparatus and an image processing method.

  There is a need for a system that provides detailed traffic information on vehicles on the road for smooth traffic or traffic accident suppression, and obtained by imaging a predetermined area including the road with an imaging device such as a video camera. There is a system that detects the presence of a vehicle by comparing a captured image with a captured image when the vehicle is not passing. However, at night, since the light from the vehicle light (headlight or taillight) of the vehicle is reflected on the road surface, the reflected image reflected on the road surface may be erroneously detected as a vehicle.

  Therefore, when a vehicle is detected by extracting a vehicle lamp as a feature point from a captured image of a vehicle traveling on the road at night or the like, a distribution pattern of a difference value between the luminance value of the input image pattern and the road surface reference luminance value is set. There has been proposed an image processing type traffic flow measuring apparatus that removes a road surface reflection image by a car lamp by making a determination (see Patent Document 1).

Further, it is determined whether or not there is road surface reflection from an image obtained by imaging a vehicle traveling on the road. When it is determined that there is road surface reflection, the luminance in the horizontal direction is set for each pixel of the processed image obtained by performing edge processing on the image. to add. The added value tends to increase at a position corresponding to the vehicle and its reflected image, and conversely tends to decrease at a position corresponding to the boundary between the vehicle and its reflected image. A vehicle detection method that can accurately detect the position of a vehicle has been proposed (see Patent Document 2).
JP-A-3-286398 JP 2003-296878 A

  However, in the image processing type traffic flow measuring device disclosed in Patent Document 1, scanning is performed for each measurement line in a predetermined direction, and when the length of the pair of luminance shift portions in the measurement line direction is larger than a certain length, the traveling After detecting the vehicle light portion of the vehicle, the presence or absence of a bright portion in front of the detected vehicle light is determined, and when the bright portion satisfies a predetermined condition, it is removed as a road surface reflection image. For this reason, when the vehicle light portion of the traveling vehicle cannot be accurately detected, the road surface reflection image cannot be detected, and the road surface reflection image may not necessarily be accurately removed.

  Further, in the vehicle detection method of Patent Document 2, when there is road surface reflection in rainy weather, the captured image is edge-processed to clarify the outline of the vehicle and its reflection image, and then each of the vehicle and its reflection image is determined. The luminance of the pixels is added in the horizontal direction, and one of the vehicles is detected based on the minimum value between the two peaks of the added value. Like a vehicle traveling at night, a vehicle using edge processing If this is difficult to detect, there is a possibility that it cannot be detected accurately.

  The present invention has been made in view of such circumstances, and a calculation unit that calculates an index value indicating the shape of the extracted vehicle lamp candidate region based on at least the luminance or hue of each pixel of the captured image, and calculation By providing a determination means for determining the presence of a vehicle light based on the index value thus determined, it is possible to determine the vehicle light and road surface reflection of a vehicle traveling at night and accurately determine the presence of the vehicle light. An object is to provide an image processing apparatus and an image processing method.

  Another object of the present invention is that the index value is a similarity to the circular shape of the vehicle lamp candidate area, so that the road surface reflection state changes depending on the road surface state, the traveling vehicle state, and the like. Even in such a case, an object of the present invention is to provide an image processing apparatus that can reliably determine the presence of a vehicle lamp by removing road surface reflection.

  Another object of the present invention is to determine the presence of a vehicle lamp based on a first threshold in an area on a captured image corresponding to an imaging area at a long distance, and to determine whether the determination means is at a short distance. In the area on the captured image corresponding to the imaging area, the vehicle lamp is determined based on the second threshold value that is larger than the first threshold value, so that the vehicle lamp can be accurately detected regardless of the distance to the vehicle. An object of the present invention is to provide an image processing apparatus capable of determining existence.

  Another object of the present invention is to provide a detection means for detecting a vehicle based on coordinates on a captured image of a vehicle lamp area corresponding to the vehicle lamp when the presence of the vehicle lamp is determined. An object of the present invention is to provide an image processing apparatus capable of accurately detecting a vehicle even at night, as distinguished from road surface reflection.

  An image processing apparatus according to a first aspect of the present invention is an image processing apparatus that performs processing for determining the presence of a vehicle lamp based on a captured image obtained by capturing an area including a road. Extraction means for extracting a vehicle lamp candidate area based on at least luminance or hue, calculation means for calculating an index value indicating the shape of the extracted vehicle lamp candidate area, and vehicle light based on the calculated index value And determining means for determining the presence of.

  An image processing apparatus according to a second invention is characterized in that, in the first invention, the index value is a similarity to a circular shape of the shape of the candidate lamp region.

  In the image processing apparatus according to a third aspect of the present invention, in the first or second aspect of the invention, the determining means determines whether the vehicle lamp The vehicle light is determined based on a second threshold value that is greater than the first threshold value in a region on the captured image corresponding to the imaging region at a short distance. .

  An image processing apparatus according to a fourth aspect of the present invention is the image processing apparatus according to any one of the first to third aspects of the present invention, on the captured image of the vehicle lamp area corresponding to the vehicle lamp when the determination means determines the presence of the vehicle lamp A detection means for detecting the vehicle based on the coordinates is provided.

  An image processing method according to a fifth aspect of the present invention is an image processing method for performing processing for determining the presence of a vehicle lamp based on a captured image obtained by capturing an area including a road. Extracting a vehicle lamp candidate area based on at least luminance or hue, calculating an index value indicating the shape of the extracted vehicle lamp candidate area, and determining the presence of a vehicle lamp based on the calculated index value It is characterized by.

  In the first invention and the fifth invention, the extraction means extracts the vehicle lamp candidate area based on at least the luminance or the hue of each pixel of the captured image obtained by imaging the area including the road. For example, when a vehicle traveling on a road at night is imaged, the captured image is based on a predetermined luminance threshold by utilizing the fact that the luminance of the vehicle's vehicle light part and the road surface reflection part by the vehicle light is relatively high. Binarization is performed, and in the binarized image, a connected region is specified and a vehicle lamp candidate region is extracted. The calculation means calculates an index value indicating the shape of the extracted vehicle lamp candidate area. The determination means determines the presence of the vehicle lamp based on the calculated index value.

  In the second invention, the index value is a similarity to the circular shape of the shape of the vehicle lamp candidate area. For example, when the circularity is used as the similarity, when the vehicle lamp candidate area is similar to a circle, it can be determined that the vehicle lamp candidate area is a vehicle lamp. Further, when the vehicle lamp candidate area is not similar to a circle, the vehicle lamp candidate area can be removed as being road surface reflection. Thus, even when the road surface reflection state (for example, the shape on the captured image) changes according to the road surface state, the state of the traveling vehicle (for example, the number of vehicles, the vehicle traveling in the adjacent lane), etc. Remove reflections.

  In the third invention, the determination means determines the presence of the vehicle lamp based on the first threshold in the area on the captured image corresponding to the imaging area at a long distance, and the imaging area at a short distance. In the region on the captured image corresponding to, the vehicle light is determined based on a second threshold value that is larger than the first threshold value. When the vehicle is at a short distance, the car light approximates a circle, and when the vehicle is at a long distance, the car light is slightly elliptical. By changing the threshold value for determining the presence of the vehicle light according to the distance, the vehicle light is determined with higher accuracy even if the shape of the vehicle light candidate region on the captured image changes.

  In the fourth invention, when the presence of the vehicle light is determined, the detection means detects the vehicle based on the coordinates on the captured image of the vehicle light region corresponding to the vehicle light. For example, when the presence of a plurality of vehicle lamps is determined, the distance between the abscissas of a pair of vehicle lamp regions that approximate the ordinate on the captured image is within a predetermined range. Detect as a vehicle.

  In the first invention and the fifth invention, calculation means for calculating an index value indicating the shape of the extracted vehicle lamp candidate area based on at least the luminance or the hue of each pixel of the captured image, and the calculated index value Therefore, the presence of the vehicle light can be determined with high accuracy by determining the vehicle light and the road surface reflection of the vehicle traveling at night. Thereby, the road surface reflection by a vehicle light can be removed and it becomes possible to apply to nighttime traffic flow measurement.

  In the second aspect of the invention, the index value is a similarity to the circular shape of the vehicle lamp candidate area, so that even when the road surface reflection state changes depending on the road surface condition, the traveling vehicle condition, etc. The presence of the vehicle light can be reliably determined by removing the road surface reflection.

  In the third invention, the determination means determines the presence of the vehicle lamp based on the first threshold in the area on the captured image corresponding to the imaging area at a long distance, and the imaging area at a short distance. In the region on the captured image corresponding to, the vehicle light is determined based on the second threshold value that is larger than the first threshold value, so that the presence of the vehicle light can be accurately detected regardless of the distance to the vehicle. Can be determined.

  In the fourth invention, when it is determined that the vehicle lamp is present, the vehicle is detected at night by providing a detection means for detecting the vehicle based on the coordinates on the captured image of the vehicle lamp area corresponding to the vehicle lamp. Even if it exists, a vehicle can be detected with high accuracy in distinction from road surface reflection.

  Hereinafter, the present invention will be described with reference to the drawings illustrating embodiments. FIG. 1 is a schematic diagram showing an outline of traffic flow measurement provided with an image processing apparatus according to the present invention. In the figure, reference numeral 1 denotes a video camera. For example, the video camera 1 has an optical axis arranged along the road direction at a position of about 10 m on the road near the intersection, and images a traffic flow in a determination area of several tens of meters on the road upstream from the intersection.

  An image processing device 2 is connected to the video camera 1, and the image processing device 2 processes a captured image obtained by imaging with the video camera 1, and the presence of a vehicle lamp (headlight or the like) of the vehicle. Make a decision. A signal control device 3 is connected to the image processing device 2 via a communication line 5, and a signal device 4 is connected to the signal control device 3.

  The image processing device 2 outputs a determination signal to the signal control device 3 when determining the presence of the vehicle lamp in the determination region. When the signal control device 3 receives the determination signal of the vehicle light from the image processing device 2, the signal control device 3 changes the right turn signal for turning the right turn signal of the traffic light 4 to blue, or changes to a blue light when the traffic light 4 is blinking red. A control signal for control such as recall sensitivity is output to the traffic light 4. In the figure, an example of three lanes on one side is shown, but the road where the video camera 1, the image processing device 2, etc. are installed is not limited to this.

  FIG. 2 is a block diagram showing the configuration of the image processing apparatus 2. The video camera 1 outputs a captured image obtained by imaging to the image input unit 21 as a video signal (analog signal). The image input unit 21 outputs the acquired video signal to the A / D conversion unit 22, and the A / D conversion unit 22 converts the input video signal into a digital signal, and uses the converted digital signal as image data. Store in the image memory 23. In this case, the image data is stored in the image memory 23 in units of one frame (for example, 640 × 480 pixels) in synchronization with the frame rate of the video camera 1 (interval at the time of imaging, for example, 30 frames per second). The

  The image processing unit 24 is constituted by an ASIC or the like, reads image data in units of one frame from the image memory 23, and binarizes the read image data based on a predetermined luminance threshold (for example, luminance) When the value is expressed by 8 bits, the luminance value “255” and the luminance value “0” are binarized). When a vehicle traveling on a road is imaged at night, the vehicle lamp (headlight) of the vehicle has the highest luminance value, and then the luminance of the road surface reflection portion by the vehicle lamp is high, making it difficult to distinguish between the two. On the other hand, the brightness value is low on the vehicle portion excluding the vehicle lights or on the road surface and surroundings where the light of the vehicle lights is not irradiated. For this reason, the brightness threshold value can be set so that the vehicle light and the road surface reflection can be distinguished from other portions. In addition, it is not limited to a brightness | luminance, A hue etc. can also be utilized.

  The image processing unit 24 performs a so-called labeling process on the binarized image, connects adjacent regions with the luminance value “255”, and extracts the connected region as a vehicle lamp candidate region.

  The image processing unit 24 calculates a circularity indicating how much the vehicle lamp candidate area is circular with respect to the extracted vehicle lamp candidate area. When a plurality of candidate vehicle lamp areas are extracted from one captured image, the circularity is calculated for each candidate lamp area. In addition to circularity as an index value (a feature value for analyzing the shape) indicating how close (similar) the shape of the candidate lamp area is to a circle, other indicators such as the degree of roundness and acicularity It is also possible to use a value.

  If the calculated circularity is greater than or equal to a predetermined threshold, the image processing unit 24 identifies the vehicle lamp candidate area as the vehicle lamp area, determines that the vehicle lamp exists in the determination area, and outputs a determination signal. In addition, when the calculated circularity is smaller than a predetermined threshold, the image processing unit 24 identifies the vehicle lamp candidate area as a road surface reflection area, and sets the luminance value of the area determined as the road surface reflection area to “0”. Thus, the road surface reflection area is removed. Thereby, the presence of the vehicle light can be accurately determined by discriminating the vehicle light and the road surface reflection.

  When the presence of the vehicle lamp is determined, the image processing unit 24 specifies the coordinates of the vehicle lamp area on the captured image. When the horizontal and vertical directions on the captured image are the X and Y axis directions, respectively, a pair of vehicle lamp areas in which the distance (interval) of the X coordinate is within a predetermined range among the vehicle lamp areas approximated by the Y coordinate. Is detected as one vehicle. Thus, the vehicle can be detected based on the presence of the vehicle light. Further, the vehicle lamp area other than the pair of vehicle lamp areas detected as a vehicle can be detected as a two-wheeled vehicle, for example. The predetermined range for detecting the vehicle can be set as appropriate according to the vehicle type (large vehicle, ordinary vehicle, etc.).

  The interface unit 25 has an interface function for transmitting a vehicle lamp presence determination signal determined by the image processing device 2 to the signal control device 3 via the communication line 5. The interface unit 25 also has an interface function for outputting vehicle detection results and the like.

3 and 4 are conceptual diagrams showing a method for calculating the circularity. The circularity is an index value indicating how close a certain area is to a circular shape. When the area of a certain region is S and the peripheral length of the region is L, the circularity E is expressed by E = 4πS / L ² . In this case, 0 <E ≦ 1. As shown in FIG. 3A, the area S of the region A in the captured image is represented by the number of pixels of the region A. Further, the peripheral length L of the region A is expressed by, for example, the distance of a line connecting the centers of the pixels that are the outer periphery of the region A.

  Next, a method for calculating the area S and the peripheral length L of the region A will be described. For the sake of simplicity, as shown in FIG. 3B, description will be made based on an enlarged image (a pixel block including 7 × 7 = 49 pixels) obtained by enlarging a partial area B of the area A. In the enlarged image, a region with a pattern is a partial region A1 of the region A, and a white region is a region other than the region A.

  As shown in FIG. 4A, the area S1 of the region A1 is “31” which is the number of pixels of the region A1. As shown in FIG. 4B, the peripheral length L1 of the region A1 is the distance of a line connecting the centers of the pixels that are the outer periphery of the region A1. Since it is a digital image, if the distance between adjacent pixels in the vertical and horizontal directions is “1.0”, and the diagonal distance between adjacent pixels is “1.5”, the peripheral length L1 is “8”.

  By applying the above-described area and perimeter calculation method to the region A, for example, assuming that the area S of the region A is “2962” and the perimeter L is “208”, the circularity E of the region A is “0.86”.

  The area of the vehicle lamp candidate area on the captured image is represented by the number of pixels of the vehicle lamp candidate area. The peripheral length of the vehicle lamp candidate area can be calculated as, for example, the distance of a line connecting the centers of the pixels that are the outer periphery of the vehicle lamp candidate area.

  The index value for determining the vehicle light is not limited to the circularity, but the degree of roundness represented by (absolute maximum length squared / area) × (π / 4), (pattern The acicular degree represented by (width / absolute maximum length) can also be used. Here, the absolute maximum length is the maximum value of the distance between any two points on the contour line of the region, and the pattern width is the distance between two straight lines that sandwich the pattern region in the direction of the absolute maximum length.

  FIG. 5 is an explanatory diagram showing the relationship between the shape of the region and the circularity. As shown in the figure, the circularity is 1.00 in the case of a perfect circle, the circularity is 0.85 in the case of an ellipse close to a circle, and the circularity is 0.71 in the case of a rectangle. Thereby, it can be seen that, for example, by setting the circularity threshold to about 0.8, it is possible to accurately discriminate between the vehicle lamp and the road surface reflection. In addition, even when the road surface reflection state changes depending on the number of traveling vehicles, road surface conditions, road surrounding conditions, etc., and the shape of the road surface reflection image on the captured image changes, the circularity (similarity to the circular shape) By using it, it is possible to reliably recognize road surface reflections from vehicle lights.

  When the vehicle is imaged, the shape of the vehicle light of the vehicle at a short distance is nearly circular, and the shape of the vehicle light of the vehicle at a long distance is elliptical. Therefore, a threshold value for determining whether or not a vehicle lamp is based on the circularity is set to a small threshold value in a region corresponding to a long distance of a captured image, and large in a region corresponding to a short distance of the captured image. By using the threshold value, it is possible to determine the vehicle light with higher accuracy regardless of the distance to the vehicle. In this case, for example, based on the vertical direction on the captured image, that is, the value of the Y coordinate, a region using a large threshold value and a region using a small threshold value can be separated, and three or more values differ stepwise. A threshold may be used. Furthermore, a predetermined initial threshold value can be set in advance, and the threshold value can be calculated by adding a coefficient to the initial threshold value based on the Y coordinate value of the vehicle lamp candidate area.

  Next, the operation of the image processing apparatus according to the present invention will be described. FIG. 6 is a flowchart showing a procedure for determining the presence of a vehicle light. The image processing unit 24 acquires a captured image (S11), and binarizes the acquired captured image (S12). In the binarized image, the image processing unit 24 connects adjacent regions by labeling processing, and extracts a region having a certain area or more as a vehicle lamp candidate region (S13). Thereby, when extracting a vehicle lamp candidate area | region, what has a small area of a vehicle lamp candidate area | region is removed as noise.

  The image processing unit 24 calculates the circularity of the extracted vehicle lamp candidate area (S14), and determines whether the calculated circularity is equal to or greater than a threshold value (S15). When the circularity is equal to or greater than the threshold value (YES in S15), the image processing unit 24 identifies the vehicle lamp candidate area as the vehicle lamp area (S16), and outputs a vehicle lamp presence determination signal (S17). It is determined whether or not the circularity is calculated for the vehicle lamp candidate area (S18).

  When the circularity is smaller than the threshold value (NO in S15), the image processing unit 24 removes the vehicle lamp candidate region from the captured image as being road surface reflection (S19), and continues the processing from step S18. When the circularity has not been calculated for all the vehicle lamp candidate regions (NO in S18), the image processing unit 24 continues the processing from step S14.

  When the circularity is calculated for all the vehicle lamp candidate areas (YES in S18), the image processing unit 24 calculates the coordinates on the captured image of the vehicle lamp area (S20), and the distance between the pair of vehicle lamp areas ( For example, it is determined whether or not the X coordinate distance on the captured image is within a predetermined range (S21). When the interval between the vehicle lamp areas is within the predetermined range (YES in S21), the image processing unit 24 detects the presence of the vehicle (S22) and determines whether or not the process has been performed for all the vehicle lamp areas ( S23).

  If the interval between the vehicle lamp areas is not within the predetermined range (NO in S21), the image processing unit 24 continues the process from step S23. When the process is not performed for all the vehicle lamp areas (NO in S23), the image processing unit 24 continues the process after step S21. If processing has been performed for all the vehicle lamp areas (YES in S23), the image processing unit 24 determines whether or not there is a processing end request (S24). If there is no processing end request (NO in S24), step S11 and subsequent steps. If there is a process end request (YES in S24), the process ends.

  FIG. 7 is a schematic diagram illustrating an example of a captured image. As shown in the figure, for example, when a vehicle traveling at night is imaged, a vehicle lamp (headlight) portion, a portion of road surface reflection caused by reflection of light from the headlight on the road surface, a contour of the vehicle, and the like are displayed.

  FIG. 8 is a schematic diagram illustrating an example of a binarized image obtained by binarizing a captured image. By binarizing the captured image shown in FIG. 7 using a predetermined luminance threshold, it is possible to extract portions having high luminance values, such as car lamps and road surface reflections, as car lamp candidate areas B, C, and D. In the figure, a region with a pattern is a region with high luminance (for example, luminance value “255”), and the other region is a region with low luminance (for example, luminance value “0”).

  FIG. 9 is a schematic diagram illustrating an example of a determination image in which the presence of a vehicle lamp is determined. The circularity of the vehicle lamp candidate areas B, C, and D in the binarized image of FIG. 8 is calculated, and the calculated circularity is compared with a predetermined threshold value, thereby removing the vehicle lamp candidate area D as an area due to road surface reflection. (Luminance value is set to “0”), and only the vehicle lamp area is extracted.

  As described above, according to the present invention, it is possible to determine the presence of a vehicle light with high accuracy by determining the vehicle light and road surface reflection of a vehicle traveling at night. Thereby, the road surface reflection by a vehicle light can be removed and it becomes possible to apply to nighttime traffic flow measurement. Further, even when the road surface reflection state changes according to the road surface state, the traveling vehicle state, and the like, the road surface reflection can be removed to reliably determine the presence of the vehicle light. In addition, it is possible to accurately determine the presence of a vehicle light regardless of the distance to the vehicle.

  In the above-described embodiment, when a vehicle lamp is present in the determination area on the road, the determination signal is output and the traffic light is controlled. However, the present invention is not limited to this. It is possible to apply to other examples. For example, the presence of vehicle lights in a predetermined determination area is continuously determined for a predetermined time, and the traffic flow is determined by performing statistical processing (for example, time average) related to the determined number of vehicle lights. It can also be used as information.

  In the above-described embodiment, as an example of traffic signal control, right turn sensitivity, recall sensitivity, and the like have been described. However, traffic signal control is not limited to this, and depending on traffic volume, It can be used for various controls such as control of lighting times and lighting intervals of blue, yellow, and red.

  The present invention can be applied not only at night, but also in daytime or in situations such as rainy weather.

It is a schematic diagram which shows the outline | summary of the traffic flow measurement provided with the image processing apparatus which concerns on this invention. It is a block diagram which shows the structure of an image processing apparatus. It is a conceptual diagram which shows the calculation method of circularity. It is a conceptual diagram which shows the calculation method of circularity. It is explanatory drawing which shows the relationship between the shape of an area | region, and circularity. It is a flowchart which shows the determination processing procedure of the presence of a vehicle light. It is a schematic diagram which shows the example of a captured image. It is a schematic diagram which shows the example of the binarized image which binarized the captured image. It is a schematic diagram which shows the example of the determination image which determined presence of the vehicle light.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Video camera 2 Image processing apparatus 3 Signal control apparatus 4 Traffic light 5 Communication line 21 Image input part 22 A / D conversion part 23 Image memory 24 Image processing part 25 Interface part

Claims (5)

In an image processing apparatus that performs processing for determining the presence of a vehicle lamp based on a captured image obtained by imaging a region including a road,
Extraction means for extracting a vehicle lamp candidate region based on at least the luminance or hue of each pixel of the captured image;
Calculating means for calculating an index value indicating the shape of the extracted vehicle lamp candidate area;
An image processing apparatus comprising: determination means for determining the presence of a vehicle light based on the calculated index value.   The image processing apparatus according to claim 1, wherein the index value is a similarity to a circular shape of a shape of the candidate lamp region. The determination means includes
In the area on the captured image corresponding to the imaged area at a long distance, the presence of the vehicle lamp is determined based on the first threshold value, and the first on the area on the imaged image corresponding to the imaged area at the short distance. The image processing apparatus according to claim 1, wherein the vehicle light is determined based on a second threshold value that is larger than the threshold value.   2. The apparatus according to claim 1, further comprising a detection unit configured to detect a vehicle based on coordinates on a captured image of a vehicle lamp region corresponding to the vehicle lamp when the determination unit determines the presence of the vehicle lamp. The image processing apparatus according to claim 3. In an image processing method for performing processing for determining the presence of a vehicle lamp based on a captured image obtained by imaging a region including a road,
Based on at least the luminance or hue of each pixel of the captured image, extract the vehicle lamp candidate area,
Calculate an index value that indicates the shape of the extracted car lamp candidate area,
An image processing method, wherein presence of a vehicle light is determined based on a calculated index value.

Images