Abstract: An edge image generating unit 12 generates an edge image, into which edge points whose variation in luminance relative to a surrounding portion is equal to or more than a predetermined value have been extracted, from an image captured by a camera 2. A candidate image portion extracting unit 13 extracts a candidate image portion, which is a candidate for an image portion of a lane mark, from the edge image. A lane mark recognizing unit 14 detects a lane mark candidate on the basis of the candidate image portion and, in the case of having detected a plurality of lane mark candidates 51b to 54b and 61b to 63b adjacent to each other, compares the luminance of a corresponding area in the captured image for each of the lane mark candidates and recognizes the lane mark candidate 51b to 54b having the maximum luminance value as a lane mark.

Abstract: This road environment recognition device is equipped with: a connecting line determination unit, which determines primary connecting lines indicating lane marks by setting candidate points at a position near a vehicle as starting points, and sequentially connecting two or more candidate points along a first direction, which is a direction away from the vehicle; and branch point determination unit, which determines whether a branch point exists on the primary connecting lines on the basis of the positional relationships between the primary connecting lines and a residual candidate point that does not form part of the primary connecting lines.

Abstract: A lane mark recognition device detects an edge located in a proximal area in front of the vehicle, and determines a lane mark candidate on the basis of the detected edge and an edge that is located in a distal area farther than the proximal area in front of the vehicle and that is continuous with the edge. Therefore, an edge of another vehicle (leading vehicle) or the like located in the distal area in a captured image is not detected as a lane mark candidate (is excluded as a non lane mark candidate).

Abstract: A road environment recognizing apparatus includes an edge image generator, a first area extractor, and an object recognizer. The edge image generating unit generates an edge image by extracting edge points from an image of a road captured by a camera. The edge points have amounts of change in luminance from surroundings, which are higher than or equal to a predetermined value. The first area extractor extracts a first area from the edge image. The first area is partitioned by the edge points and has a luminance higher than the luminances of the surroundings. If a second area having a luminance lower than the luminances of the surroundings exists around the first area in the edge image, the object recognizer recognizes a projection on the road on the basis of a third area resulting from joining of the first area and the second area.

Abstract: Provided is an image processing device in which a region targeted for edge extraction of a taken image is divided into a plurality of partial regions 11(1) to 11(7) for which threshold values for edge extraction different from each other are set, respectively. An edge extraction unit 5, which extracts edges from the region targeted for edge extraction, performs processing of extracting the edges in the partial regions 11(1) to 11(7) by using the threshold values set so as to correspond to the each of the partial regions 11(1) to 11(7).

Abstract: An edge image generating unit 12 generates an edge image 22, into which edge points whose variation in luminance relative to a surrounding portion is equal to or more than a predetermined value have been extracted, from an image 21 captured by a camera 2. A candidate image portion extracting unit 13 extracts a candidate image portion, which is a candidate for an image portion of a dashed-line lane mark, from the edge image 22. A dashed-line lane mark recognizing unit 14 recognizes the dashed-line lane mark based on a plurality of candidate image portions each of whose length in the corresponding real space is within a predetermined range and which are continuous in a given direction among the candidate image portions.

Abstract: An edge image generating unit 12 generates an edge image, into which edge points whose variation in luminance relative to a surrounding portion is equal to or more than a predetermined value have been extracted, from an image captured by a camera 2. A candidate image portion extracting unit 13 extracts a candidate image portion, which is a candidate for an image portion of a lane mark, from the edge image. A lane mark recognizing unit 14 detects a lane mark candidate on the basis of the candidate image portion and, in the case of having detected a plurality of lane mark candidates 51b to 54b and 61b to 63b adjacent to each other, compares the luminance of a corresponding area in the captured image for each of the lane mark candidates and recognizes the lane mark candidate 51b to 54b having the maximum luminance value as a lane mark.

Abstract: This road environment recognition device is equipped with: a connecting line determination unit, which determines primary connecting lines indicating lane marks by setting candidate points at a position near a vehicle as starting points, and sequentially connecting two or more candidate points along a first direction, which is a direction away from the vehicle; and branch point determination unit, which determines whether a branch point exists on the primary connecting lines on the basis of the positional relationships between the primary connecting lines and a residual candidate point that does not form part of the primary connecting lines.

Abstract: Provided is a lane recognition device capable of extracting linear elements derived from lane marks from a linear element extraction image obtained by processing a captured image and recognizing lane boundary lines. Local areas 47 are set for a lane extraction area 45 set in a linear element extraction image 40 that each of linear elements is included in one or a plurality of the local areas 47, having a predetermined size, and a local straight line 44 of each local area is determined (vx) and (?) associated with the direction and the intersection x with a predetermined reference horizontal line 46 are calculated for each local straight line 44. Each local straight line 44 is defined as one vote, being casted to (vx, ?) of a voting space 56. Lane boundary lines are recognized from detection straight lines, whose direction and the intersection x determined based on vote results.

Abstract: A vehicle periphery monitoring device includes: a first edge image generation element 5 which generates a first edge image on the basis of luminance components of a captured image acquired by an in-vehicle camera 2; a second edge image generation element 6 which generates a second edge image on the basis of hue components or saturation components of the captured image; a composite edge image generation element 7 which generates a composite edge image formed by combining the first edge image and the second edge image; and an object classification identification element 8 which identifies whether or not the object is a prescribed kind of structure on the basis of the external shape of the object represented by the composite edge image.

Abstract: A vehicular image processing device is provided with a kernel setting unit which sets a plurality of smoothing kernels which have a width supposed to be between the width of the lane mark and that of the road to the image acquired from a photographing unit, a smoothing unit which smoothes the acquired image by filtering using the set smoothing kernels, a variation degree calculating unit which calculates the variation degree of the pixel value of each pixel in the acquired image with respect to the smoothed image, and a pixel value replacing unit which replaces the pixel value of a pixel which is of the acquired image and has the variation degree not greater than a predefined value with a specific value.

Abstract: Provided is a lane recognition device capable of extracting linear elements derived from lane marks from a linear element extraction image obtained by processing a captured image and recognizing lane boundary lines. Local areas 47 are set for a lane extraction area 45 set in a linear element extraction image 40 that each of linear elements is included in one or a plurality of the local areas 47, having a predetermined size, and a local straight line 44 of each local area is determined (vx) and (?) associated with the direction and the intersection x with a predetermined reference horizontal line 46 are calculated for each local straight line 44. Each local straight line 44 is defined as one vote, being casted to (vx, ?) of a voting space 56. Lane boundary lines are recognized from detection straight lines, whose direction and the intersection x determined based on vote results.

Abstract: A vehicle periphery monitoring device includes: a first edge image generation element 5 which generates a first edge image on the basis of luminance components of a captured image acquired by an in-vehicle camera 2; a second edge image generation element 6 which generates a second edge image on the basis of hue components or saturation components of the captured image; a composite edge image generation element 7 which generates a composite edge image formed by combining the first edge image and the second edge image; and an object classification identification element 8 which identifies whether or not the object is a prescribed kind of structure on the basis of the external shape of the object represented by the composite edge image.

Abstract: A lane recognition apparatus for the vehicle includes: a lane mark detection image acquisition device which acquires an image divided into a first image region composed of pixels having pixel values supposed to be image portions of a lane mark defining the lane and a second image region composed of other pixels from a color image of a road; a clustering device which divides the first image region into subregions each composed of adjacent pixels; a color determination device which determines a color of the subregions; a lane recognition device which performs line component extraction for each group, considering the subregions determined to have the same color as a single group, and recognizes a line position of the lane defined by the lane mark from the extracted line component; and a vehicle equipment control device which controls equipment of a subject vehicle according to the lane recognition result.

Abstract: A vehicular image processing device is provided with a kernel setting unit which sets a plurality of smoothing kernels which have a width supposed to be between the width of the lane mark and that of the road to the image acquired from a photographing unit, a smoothing unit which smoothes the acquired image by filtering using the set smoothing kernels, a variation degree calculating unit which calculates the variation degree of the pixel value of each pixel in the acquired image with respect to the smoothed image, and a pixel value replacing unit which replaces the pixel value of a pixel which is of the acquired image and has the variation degree not greater than a predefined value with a specific value.

Abstract: A lane recognition apparatus for the vehicle includes: a lane mark detection image acquisition device which acquires an image divided into a first image region composed of pixels having pixel values supposed to be image portions of a lane mark defining the lane and a second image region composed of other pixels from a color image of a road; a clustering device which divides the first image region into subregions each composed of adjacent pixels; a color determination device which determines a color of the subregions; a lane recognition device which performs line component extraction for each group, considering the subregions determined to have the same color as a single group, and recognizes a line position of the lane defined by the lane mark from the extracted line component; and a vehicle equipment control device which controls equipment of a subject vehicle according to the lane recognition result.

Abstract: Roadside three-dimensional objects arranged along the road at irregular intervals are used to accurately estimate a shape of the road. A three-dimensional object extractor extracts a three-dimensional object existing in an image captured by an image pickup device. A feature point extractor extracts a feature point from the three-dimensional object and a feature point corrector corrects a displacement of the feature point according to a detected running state of the vehicle. A residual image generator generates a residual image of the feature point whose displacement is corrected and the running state of the subject vehicle. A vanishing point calculator calculates a vanishing point and a road shape estimator estimates a road shape based on a straight line closest to both right and left sides of the subject vehicle from the straight lines passing through the vanishing point.