Abstract: An object detection apparatus for detecting an object from an image obtained by taking a front view picture of a road in a traveling direction of a vehicle includes a camera unit for taking the front view picture of the road and inputting the image; a dictionary modeling the object; a search unit for searching the image with a search window; a histogram production unit for producing a histogram by comparing the image in the search window with the dictionary and counting a detection frequency in a direction parallel to a road plane; and a detection unit for detecting the detection object by detecting a unimodal distribution from the histogram.

Abstract: A reflection-point estimation section constituting an object detection apparatus first calculates a rotational correction angle for each estimated reflection point Pn by using the following equation, using the fact that the ratio of the amount of change in a detected distance dn to a movement amount In of a sonar is nearly equal to a tangent with respect to the direction normal to the surface of an object: Rotational correction angle?=a tan {(dn+1?dn)/In} in which n is a natural number. Subsequently, the reflection-point estimation section obtains estimated reflection points Pn? after correction, by rotating each estimated reflection point Pn based on the calculation result.

Abstract: An object detection apparatus for detecting an object from an image obtained by taking a front view picture of a road in a traveling direction of a vehicle includes a camera unit for taking the front view picture of the road and inputting the image; a dictionary modeling the object; a search unit for searching the image with a search window; a histogram production unit for producing a histogram by comparing the image in the search window with the dictionary and counting a detection frequency in a direction parallel to a road plane; and a detection unit for detecting the detection object by detecting a unimodal distribution from the histogram.

Abstract: A point-sequence approximation apparatus determines whether a proportion of a straight-line approximation region is high as compared to that of a curve approximation region. In addition, if it is determined that the proportion of the straight-line approximation region is high as compared to that of the curve approximation region, the point-sequence approximation apparatus repeatedly performs random setting of a candidate of an approximation straight line, and evaluation of approximation accuracy of the candidate of the approximation straight line based on distances between the candidate of the approximation straight line and the points constituting the point sequence. Moreover, the point-sequence approximation apparatus performs straight-line approximation by using the candidate of the approximation straight line with the best result of evaluation of the approximation accuracy as an approximation straight line.

Abstract: A near-infrared night vision device to which a pedestrian detection device is applied includes a near-infrared projector, a near-infrared camera, a display and an ECU. By executing programs, the ECU constitutes a pedestrian candidate extraction portion and a determination portion. The pedestrian candidate extraction portion extracts pedestrian candidate regions from near-infrared images. The determination portion normalizes the sizes and the brightnesses of the pedestrian candidates extracted by the pedestrian candidate extraction portion, and then computes the degrees of similarity between the normalized pedestrian candidates. The determination portion determines that a pedestrian candidate having two or more other pedestrian candidates whose degree of similarity with the pedestrian candidate is greater than or equal to a predetermined value is not a pedestrian.

Abstract: A reflection-point estimation section constituting an object detection apparatus first calculates a rotational correction angle for each estimated reflection point Pn by using the following equation, using the fact that the ratio of the amount of change in a detected distance dn to a movement amount In of a sonar is nearly equal to a tangent with respect to the direction normal to the surface of an object: Rotational correction angle?=atan{(dn+1?dn)/In} in which n is a natural number. Subsequently, the reflection-point estimation section obtains estimated reflection points Pn? after correction, by rotating each estimated reflection point Pn based on the calculation result.

Abstract: A point-sequence approximation apparatus determines whether a proportion of a straight-line approximation region is high as compared to that of a curve approximation region. In addition, if it is determined that the proportion of the straight-line approximation region is high as compared to that of the curve approximation region, the point-sequence approximation apparatus repeatedly performs random setting of a candidate of an approximation straight line, and evaluation of approximation accuracy of the candidate of the approximation straight line based on distances between the candidate of the approximation straight line and the points constituting the point sequence. Moreover, the point-sequence approximation apparatus performs straight-line approximation by using the candidate of the approximation straight line with the best result of evaluation of the approximation accuracy as an approximation straight line.