JP2009223818A - Image processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus capable of eliminating an effect of a portion depending on an environmental condition, and capable of calculating the similarity to other image, even in an image having the portion depending on the environmental condition. <P>SOLUTION: The image processing apparatus for calculating the similarity between the two images includes an edge image generating part 103 for extracting an edge from the image, and for generating an edge image, a nondependent image generating part 104 for generating an environment fluctuation nondependent image, based on the edge image in one out of the two images generated by the edge image generating part, and based on an edge image of a reference image not depending on environment fluctuation generated by the edge image generating part, and a similarity calculating means 105 for calculating the similarity between the two images, based on the environment fluctuation nondependent image generated by the nondependent image generating part. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image processing apparatus.

  Various methods have been proposed for calculating the similarity between two images. For example, in Patent Document 1, the difference between the density gradient direction of each pixel of the input image and the density gradient direction of the corresponding pixel of the model image is used as an evaluation value, and this evaluation value is weighted by the density gradient strength of the model image. A similarity calculation device that calculates a similarity by adding weighted evaluation values is disclosed.

In Patent Document 2, in order to cope with environmental changes, a normal background image is stored, an identification symbol whose similarity is to be calculated is combined with the stored image, and a pattern is formed on the combined image. A technique for performing matching is disclosed.
Japanese Patent No. 3598651 JP 2007-188221 A

  However, in Patent Document 1, when the model image and the input image are captured under different environmental conditions, there is a portion that occurs on the input image depending on environmental conditions such as a shadow, and the density gradient of this portion is If it is large, there is a problem that the similarity is calculated due to the influence of the shadow portion.

  In addition, since Patent Document 2 is not a technique for calculating the similarity between images affected by environmental conditions, an appropriate similarity is calculated for images having portions that occur depending on the environmental conditions. There is a problem that it is not possible.

  The present invention has been made in view of the above problems, and even if an image has a part depending on the environmental condition, the influence of the part depending on the environmental condition is removed to calculate the similarity with other images. An object of the present invention is to provide an image processing apparatus capable of performing the above.

  In an image processing apparatus that calculates a similarity between two images, an edge image is extracted from the image and an edge image is generated, and the two images generated by the edge image generation unit An independent image generation unit that generates an environment variation independent image based on an edge image of one of the images and an edge image of a reference image that does not depend on the environmental variation generated by the edge image generation unit; And a similarity calculation unit that calculates a similarity between the two images based on an environment variation independent image generated by the image generation unit.

  According to the present invention, even if an image has a part depending on the environmental condition, the similarity with other images can be calculated by removing the influence of the part depending on the environmental condition.

  Embodiments of the present invention will be described below with reference to the drawings.

An image processing apparatus according to an embodiment of the present invention has a functional configuration as shown in FIG.
FIG. 1 is a schematic functional block diagram of an image processing apparatus according to Embodiment 1 of the present invention.

  The image processing apparatus 100 according to the first embodiment includes a storage unit 101, an image input unit 102, an edge image generation unit 103, an independent image generation unit 104, and a similarity calculation unit 105.

  The storage unit 101 stores a reference image. Here, the reference image is an image that does not depend on environmental conditions, that is, an image in which a shadow of a building or the like is not captured and an image that should be originally captured is captured. Note that the storage unit 101 may store an image other than the reference image. At this time, the reference image and the image other than the reference image can be identified.

  In addition, the reference image may be created by manually specifying a portion that does not depend on environmental conditions among the captured images. If there are a plurality of images at the same position, the reference image may have a single edge. The fewest images may be used as reference images. The latter is based on the idea that an image with few edges does not have a portion depending on environmental conditions such as a shadow, and an image captured on a cloudy day or the like is considered to be a reference image.

  The image input unit 102 inputs two images captured by the imaging unit, outputs one of the two images to the edge image generation unit 103, and outputs the two images to the similarity calculation unit 105.

  Here, the reference image and the two images whose similarity is calculated will be described with reference to FIG. FIG. 2A shows a reference image. In the reference image, roads, roadside buildings, stop lines, and manholes are captured, and shadows that depend on environmental conditions are not captured.

  FIG. 2B is a diagram illustrating the image 1. In image 1, the shadow A of the building is captured on the road. FIG. 2C shows the image 2. In image 2, the shadow B of the building is captured on the road. In this case, the reference image in FIG. 2A is stored in the storage unit 101, the image 1 in FIG. 2B and the image 2 in FIG. 2C are captured by the imaging unit, and the two images are the image input unit. 102 is input.

  In order to calculate the similarity between the images, the image input means 102 basically inputs and outputs two images. However, when it is desired to calculate a plurality of similarities at once, a plurality of images are input. You may make it input and output.

  Here, it is assumed that an image captured by the imaging unit is input, but two images already stored in the storage unit 101 may be read out.

  Returning to FIG. 1, the edge image generation unit 103 reads a reference image from the storage unit 101, and generates an edge image by extracting an edge from the read reference image and the image acquired from the image input unit. Further, the edge image generation unit 103 outputs the generated two edge images to the independent image generation unit 104. Note that a known technique such as a Sobel filter, Laplacian, or Canny is used for edge extraction.

  The independent image generation unit 104 generates an environment variation independent image from the two edge images acquired from the edge image generation unit 103. Here, the environment variation-independent image refers to an edge image obtained by extracting a portion where the edges of two edge images are common.

  Here, the edge image and the environment variation-independent image will be described with reference to FIG. FIG. 3A shows an edge image of the reference image shown in FIG. FIG. 3B is a diagram illustrating an edge image of the image 1 illustrated in FIG.

  FIG. 3C is a diagram showing an environment variation-independent image showing an edge portion common to the edge image of the reference image shown in FIG. 3A and the edge image of the image 1 shown in FIG. That is, the environment variation-independent image is an edge image from which an influence such as a shadow depending on the environmental condition is removed.

  Returning to FIG. 1, the similarity calculation unit 105 includes a difference value calculation unit 106 and a weighting unit 107, and calculates the similarity between images.

  The difference value calculation unit 106 calculates a difference value for the two images acquired from the image input unit 102 and generates a difference image. Here, the difference value is the absolute difference value of the pixel values at corresponding positions in the two images. Although the difference image is generated here, it is not necessary to generate the difference image if the difference absolute value and the pixel position are associated and remembered.

  The difference image will be described with reference to FIG. The images shown in FIGS. 4A and 4B are the image 1 and the image 2 shown in FIG. FIG. 4C is a diagram illustrating a difference image between the image 1 illustrated in FIG. 4A and the image 2 illustrated in FIG.

  As shown in FIG. 4C, the difference image shows a difference portion between the image 1 and the image 2. In the example of FIG. 4, the difference portion between image 1 and image 2 is the difference portion between shadow A and shadow B.

  Returning to FIG. 1, the weighting unit 107 weights the difference image acquired from the difference value calculation unit 106 based on the environment variation independent image acquired from the independent image generation unit 104 and calculates the similarity.

  The similarity calculation will be specifically described with reference to FIG. The weight based on the environment variation dependent image shown in FIG. 5B is, for example, a weight coefficient of 1 is set for a pixel having an edge, and a weight coefficient of 0 or more and less than 1 is set for a pixel having no edge. . In the example shown in FIG. 5B, the weight coefficient is set to 0 for a pixel having no edge portion. As for this weighting factor, the weighting factor can be set as appropriate by increasing the weighting of the position not depending on the environmental condition and weakening the weighting of the position depending on the environmental condition.

  The degree-of-similarity calculation means 105 calculates the degree of similarity by accumulating the weighting coefficient for each pixel (difference value) of the difference image and accumulating the accumulated difference values. Thus, the similarity between images can be calculated after removing the influence of the position depending on the environmental conditions. In this case, the smaller the similarity is, the more the two images are determined to be similar.

  In the example of FIG. 5, since the difference image between the image 1 and the image 2 is greatly affected by the shadow, even if the similarity is calculated by general SAD (Sum of Absolute Differences) or the like as it is, It may be determined that the image 2 is not similar. However, by adding the weighting coefficient of FIG. 5B to the difference value and adding the difference value obtained by integrating the weighting coefficient to obtain the similarity, the influence of the shadow can be removed. Is judged to be similar.

  Further, it is assumed that a vehicle or the like that is not included in the reference image is captured in addition to the shadow that occurs in the image 1 under environmental conditions. At this time, for the vehicle, it is desirable to leave an influence on the similarity calculation. Therefore, it is necessary to distinguish between the shadow A and the vehicle. For this distinction, position information, time information, and weather information at the time of imaging are further input, and a shadow area such as a building is estimated from these information. Next, it is possible to distinguish a shadow generated by an environmental condition and a vehicle depending on whether the estimated region is similar in position and shape.

  FIG. 6 is a flowchart of similarity calculation processing in the image processing apparatus according to the first embodiment. In step 601, the image input unit 102 inputs two images picked up by the image pickup unit, and outputs one of the two input images to the edge image generation unit 103.

  Proceeding to step 602 following step 601, the edge image generation unit 103 generates an edge image for the reference image acquired from the storage unit 101 and the image acquired from the image input unit 102, and the edge image is independent. It outputs to the image generation means 104.

  Proceeding to step 603 following step 602, the independent image generation unit 104 generates an environment variation independent image based on the two edge images acquired from the edge image generation unit 103, and the environment variation independent image is determined as the similarity. The data is output to weighting means 107 included in calculation means 105.

  In step 604 following step 603, the difference value calculation unit 106 included in the similarity calculation unit 105 calculates a difference value between the two images acquired from the image input unit 102, generates a difference image, and generates a difference. The image is output to the weighting means 107.

  Proceeding to step 605 following step 604, the weighting means 107 adds the weighting factor based on the environment variation independent image acquired from the independent image generating means 104 to the difference image acquired from the difference value calculating means 106.

  Proceeding to step 606 following step 604, the similarity calculation unit 105 calculates the similarity by accumulating the difference values weighted by the weighting unit 107.

  As described above, according to the image processing apparatus according to the first embodiment, even if an image has a part depending on the environmental condition, the similarity with the other image is calculated by removing the influence of the part depending on the environmental condition. Can do. Also, when calculating the similarity between images that have parts that depend on environmental conditions, it is possible to reduce the number of images that are originally similar but are judged not to be similar due to the influence of parts that depend on environmental conditions. be able to. Furthermore, if there is a region that is really different between the two images other than the portion that depends on the environmental conditions, it can be determined that the two images are not similar under the influence of the region that is really different.

(Modification 1)
In the first modification of the image processing apparatus according to the first embodiment, an environment variation dependent image opposite to the environment variation independent image generated in the first embodiment is generated, and the similarity between the images is calculated. An environment variation dependent image is an image showing a portion such as a shadow that occurs depending on environmental conditions.

  In such a case, the environment variation-dependent image is an edge image such as a shadow generated depending on the environmental conditions, so it is desired to remove the influence of this portion. That is, by setting the weighting coefficient of the pixels surrounded by the edge of the environment variation dependent image to a small value, for example, 0, and integrating the difference image, the similarity of the region other than the shadow can be calculated.

  More specifically, instead of the independent image generating unit 104, the dependent image generating unit extracts an edge that is only in one of the edge images based on the two input edge images, and generates an edge image. The weighting means 107 sets the weighting coefficient of the part surrounded by the edge of the environment variation dependent image to 0 and integrates it to the difference image. Other processes are the same as those in the first embodiment.

  As described above, according to the image processing apparatus according to the modified example 1, it is possible to calculate the similarity between images in other regions by removing only the part depending on the environmental condition.

(Modification 2)
In the second modification of the image processing apparatus according to the first embodiment, one environment variation-independent image is generated. However, an environment variation-independent image is generated for each of the two images whose similarity is to be calculated.

  More specifically, the edge image generation unit 103 acquires two images from the image input unit 102 and generates an edge image for each of them. The independent image generation unit 104 generates the environment variation independent images for the two edge images using the edge image of the reference image.

  At this time, an image obtained by extracting only a portion common to the two environment variation independent images becomes a final environment variation independent image. That is, the final environment variation-independent image is an edge image in which a portion depending on the environmental conditions of the two images is removed.

  The subsequent processing is the same as in the first embodiment if the independent image generation unit 104 outputs the final environment variation independent image to the similarity calculation unit 105.

  As described above, according to the image processing apparatus according to the modified example 2, it is possible to calculate the similarity between images after removing the part depending on the environmental condition of the two images for calculating the similarity. Note that the image processing apparatus according to the second modification can also apply the first modification.

  In order to calculate the similarity between two images, the image processing apparatus according to the second embodiment uses the edge image of the two images to calculate the similarity between the two images. Details will be described with reference to FIG.

  FIG. 7 is a schematic functional block diagram of an image processing apparatus 700 according to the second embodiment. An image processing apparatus 700 illustrated in FIG. 7 includes a storage unit 101, an image input unit 701, an edge image generation unit 702, an independent image generation unit 703, and a similarity calculation unit 704. In the configuration shown in FIG. 7, the same configurations as those in FIG. 1 are denoted by the same reference numerals as those in FIG.

  The image input unit 701 inputs two captured images and outputs the two input images to the edge image generation unit 702. Here, the image input unit 701 does not necessarily need to input two images at the same time, and may input one image at a time in the order of image pickup by the image pickup unit.

  The edge image generation unit 702 extracts edges from the two images acquired from the image input unit 701 and the reference image read from the storage unit 101, and generates an edge image.

  In addition, the edge image generation unit 702 outputs one of the two edge images of the two images and the edge image of the reference image to the independent image generation unit 703. Further, the edge image generation unit 702 outputs the other edge image of the two edge images of the two images to the similarity calculation unit 705.

  If there is a pixel value indicating an edge at a corresponding position between the edge image of the reference image and one of the two edge images, the independent image generation unit 703 can detect the pixel value at the corresponding position of the reference image. Is set as the pixel value at the same position of the environment variation independent image.

  Further, the non-dependent image generation unit 703 sets 0 as an environment variation-independent image if there is no pixel value indicating an edge at a corresponding position between the edge image of the reference image and one of the two edge images. Are set as pixel values at the same position.

  The independent image generation unit 703 outputs the generated environment variation independent image for storage in the independent image storage unit 704.

  The independent image storage unit 704 stores the acquired environment information independent image. Although the storage unit 101 and the independent image storage unit 704 are configured separately, if the reference image and the environment variation independent information are stored separately, they may be stored in the same storage unit. .

  The similarity calculation unit 705 first determines, for each pixel, an edge variation image that is read from the independent image storage unit 704 and the other edge image of the two edge images acquired from the edge image generation unit 702 for each pixel. It is determined whether or not it has a pixel value indicating. At this time, for the position having the pixel value of 0 in the environment variation-independent image, the above determination with respect to the pixel of the other edge image is not performed. Thus, useless processing can be reduced.

  Next, as a result of the above determination, the similarity calculation unit 705 determines that if there is a pixel value indicating an edge at a corresponding position in the other edge image at a position having a non-zero pixel value in the environment variation-independent image. The pixel values of the variation-independent image are accumulated in the similarity.

  That is, the similarity in this embodiment is obtained by calculating the pixel value at the position corresponding to the reference image at a position having an edge common to all the edge images with respect to the edge image of the reference image and all the edge images of the two images. The accumulation is the similarity.

  Here, the image of the determination in the similarity calculation of the second embodiment will be described with reference to FIG. FIG. 8A is a diagram illustrating an edge image of an environment variation independent image. FIG. 8B is a diagram illustrating an example of an environment variation-independent image corresponding to the edge image illustrated in FIG. As shown in FIG. 8B, the pixel value of the reference image is set at a position having the edge of FIG. FIG. 8C is a diagram illustrating an edge image of the image 2.

  When the similarity is calculated using the images of FIGS. 8B and 8C, the above determination is made for each pixel, and the position having a non-zero pixel value of the environment variation-independent image, that is, FIG. When the edge image shown in FIG. 8C has a pixel value indicating an edge at a position having a pixel value other than 0 as shown in (), the pixel values of the environment variation-independent image are accumulated to obtain a similarity. Taking FIG. 8 as an example, similarity is 88 + 20 + 60 +... + 35 + 120. At this time, the larger the similarity value, the more similar the two images.

  In the second embodiment, as a result of the determination, the pixel values of the environment variation-independent image are accumulated to obtain the similarity, but the number of common edge positions is counted and this count value is used as the similarity. be able to.

  In addition, when the number of common edge positions is counted and the count value is used as a similarity, it is only necessary that the environment variation-independent image can identify the edge. Therefore, the pixel value with the edge is set to 1 and the pixel without the edge The value may be set to 0 to generate an environment variation independent image.

  FIG. 9 is a flowchart of similarity calculation in the image processing apparatus according to the second embodiment. In step 901, the image input unit 701 inputs two captured images and outputs the two input images to the edge image generation unit 702.

  Progressing to step 902 following step 901, the edge image generation unit 702 generates edge images for the two images acquired from the image input unit 701 and the reference image read from the storage unit 101. In addition, the edge image generation unit 702 outputs the edge image of the reference image and one of the two edge images to the independent image generation unit 703 among the generated edge images, and outputs the edge image of the two edge images. The other edge image is output to the similarity calculation means 705.

  Progressing to step 903 following step 902, the independent image generation unit 703 generates an environment variation independent image based on the edge image acquired from the edge image generation unit 702, and converts the environment variation independent image into the independent image. Output for storage in storage means 704. The generation process of the environment variation independent image will be described later with reference to FIG.

  Proceeding to step 904 following step 903, the independent image storage unit 704 stores the environment variation independent image acquired from the independent image generation unit 704.

  Proceeding to step 905 following step 904, the similarity calculation unit 705 extracts the environment variation independent image read from the independent image storage unit 704 and the other edge image acquired from the edge image generation unit 702 for each pixel. The following determination is performed and the determination result is output.

  The determination is a determination as to whether or not there is a pixel value indicating the edge of the other edge image at a corresponding position at a position having a non-zero pixel value on the environment variation independent image. The determination result is the pixel value of the environment variation-independent image when the determination result is affirmative, and is 0 when the determination result is negative. Another determination result may be 1 when the determination result is affirmative and may be 0 when the determination result is negative.

  Proceeding to step 906 following step 905, the similarity calculation means 705 accumulates the determination results and sets the accumulated value as the similarity.

  Here, the environment variation-independent image generation processing in step 903 will be described with reference to FIG. FIG. 10 is a flowchart of the environment variation independent image generation process in the image processing apparatus according to the second embodiment.

  In step 1001, the independent image generation unit 703 determines whether the pixel value is not 0 for each pixel on the edge image of the reference image. When the determination result in step 1001 is YES, the process proceeds to step 1002, and the independent image generation unit 703 corresponds to the edge image of one image at a position having a non-zero pixel value on the edge image of the reference image. It is determined whether or not the pixel value at the position has a pixel value indicating an edge.

  When the determination result in step 1002 is YES, the independent image generation unit 703 sets the pixel value of the reference image as the pixel value of the position of the environment variation independent image corresponding to the position determined as YES in step 1002. To do.

  If the determination result in step 1001 is NO, or if the determination result in step 1002 is NO, 0 is set as the pixel value of the position of the environment variation-independent image corresponding to each determined position.

  Progressing to step 1005 following step 1003 or step 1004, it is determined whether the independent image generation means 703 has performed step 1001 on all pixels. Note that the determination in step 1005 may be a determination as to whether or not all pixel values on the environment variation-independent image have been set.

  If the decision result in the step 1005 is YES, the environment variation independent image generation process is ended. If the determination result is NO, the process proceeds to step 1006, and the independent image generation unit 703 reads the next pixel on the edge image of the reference image, and returns to step 1001.

  As described above, according to the image processing apparatus according to the second embodiment, even if an image includes a part depending on the environmental condition while considering the processing load in calculating the similarity, the influence of the part depending on the environmental condition is removed. It is possible to calculate the similarity to the image. This is because the environment variation-independent image stored in the independent image storage unit 704 is used for similarity calculation, so that it is not necessary to perform the environment variation-independent image generation process each time.

  Here, FIG. 11 shows an application example of the image processing apparatus according to the embodiment or the modification described above. FIG. 11 is a diagram illustrating an example when the image processing apparatus according to the first embodiment is mounted on a vehicle. An imaging unit is provided in front of the vehicle and is connected to an image processing device mounted on the vehicle. Accordingly, when it is determined that the images taken at the same position are not similar, it is possible to notify the driver that there is an abnormal object by issuing a warning or the like.

  In such a case, for the positional deviation and rotational deviation of the captured image, the inclination of the vehicle is detected using a gyro function mounted on the vehicle, and the GPS (Global Positioning System) function of a navigation system or the like is used. By grasping the current position, it is possible to correct the deviation.

  The application example of the present invention is not limited to the above example, and it can be applied when a reference image is provided and the similarity between two captured images is to be calculated.

  Further, the configuration of the first embodiment and the configuration of the second embodiment are configured in one image processing apparatus, and the configuration of the second embodiment is adopted when the edge can be appropriately extracted. 1 may be adopted. Thus, appropriate similarity calculation can be performed according to the captured image.

  If the edge cannot be extracted properly, it is considered that the edge cannot be extracted, for example, when a blurred image is taken. Therefore, the configurations of the first embodiment and the second embodiment may be switched depending on whether the number of extracted edges is equal to or larger than a predetermined number or whether the ratio of the edge to the image is larger than a predetermined value.

  Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and other than the above-described modifications within the scope of the gist of the present invention described in the claims. In addition, various modifications and changes can be made.

1 is a diagram illustrating schematic functional blocks of an image processing apparatus according to Embodiment 1. FIG. The figure which shows the example of a reference image, the image 1, and the image 2. FIG. The figure which shows the example of the edge image of a reference image, the edge image of the image 1, and an environment variation independent image. The figure which shows the example of the difference image of the image 1 and the image 2. FIG. The figure which shows the example of the process in similarity calculation. 6 is a flowchart of similarity calculation processing in the image processing apparatus according to the first embodiment. FIG. 10 is a diagram illustrating schematic functional blocks of an image processing apparatus according to a second embodiment. The figure which shows the example of the edge image of an environmental variation independent image, the environmental variation independent image, and the edge image of the image 2. FIG. 9 is a flowchart illustrating similarity calculation processing in the image processing apparatus according to the second embodiment. The flowchart which shows an environmental variation independent image generation process. 1 is a diagram illustrating an application example of an image processing apparatus according to the present embodiment.

Explanation of symbols

100, 700 Image processing apparatus 101 Storage means 102, 701 Image input means 103, 702 Edge image generation means 104, 703 Independent image generation means 105, 705 Similarity calculation means 106 Difference value calculation means 107 Weighting means 704 Independent image storage means

Claims (8)

In an image processing apparatus that calculates the similarity between two images,
Edge image generation means for extracting an edge from an image and generating an edge image;
Based on the edge image of one of the two images generated by the edge image generation means and the edge image of the reference image independent of the environment fluctuation generated by the edge image generation means. A non-dependent image generating means for generating a dependent image;
An image processing apparatus comprising: a similarity calculation unit that calculates a similarity between the two images based on an environment variation independent image generated by the independent image generation unit. The similarity calculation means includes:
A difference value calculating means for calculating a difference value between images for each pixel between the two images;
Weighting means for weighting the difference value calculated by the difference value calculating means based on the environmental variation independent image;
The image processing apparatus according to claim 1, wherein the similarity is calculated based on a difference value weighted by the weighting unit. The independent image generation means includes
An edge image composed of a common edge with respect to the edge of the reference image extracted by the edge image generation unit and the edge of the one image extracted by the edge image generation unit is independent of the environmental variation. The image processing apparatus according to claim 2, wherein the image processing apparatus is an image. Storage means for storing the environment variation independent image generated by the independent image generation means;
The similarity calculation means includes:
Based on the environment variation-independent image stored by the storage unit and the edge image of the other image of the two images generated by the edge image generation unit, the similarity between the two images is determined. The image processing apparatus according to claim 1, wherein the image processing apparatus is calculated. The independent image generation means includes
When there is a pixel value indicating an edge at a position corresponding to the edge image of the one image and the edge image of the reference image, the pixel of the reference image is used as the pixel value of the position of the environment variation independent image A value is set, and when there is no pixel value indicating an edge at the corresponding position, the environment variation independent image is generated by setting 0 as the pixel value at the position of the environment variation independent image. The image processing apparatus according to claim 4, wherein: The similarity calculation means includes:
5. The similarity is calculated based on whether or not the environment-independent image and the edge image of the other image have a non-zero pixel value for each pixel. Or the image processing apparatus of 5. The similarity calculation means includes:
The image processing apparatus according to claim 6, wherein the determination is not performed for a position having a pixel value of 0 on the environment variation independent image. The similarity calculation means includes:
When the edge image of the other image has a pixel value indicating an edge at a position having a non-zero pixel value on the environment variation independent image, the pixel value of the environment variation independent image at the position is accumulated. The image processing apparatus according to claim 6, wherein the similarity is calculated by performing the operation.

Images