A three-dimensional shape input device for entering a three-dimensional shape of an object by a light projection method is provided, which includes a light projecting portion for projecting detection light onto an object, and an imaging portion for receiving reflected light from the object. The device includes a storage portion for storing a parameter that is used for computing the three-dimensional shape, a computing portion for computing the three-dimensional shape of the object by using image data delivered from the imaging portion and the parameter, and a dividing portion for dividing an imaging surface of an image sensor into plural areas. The storage portion stores parameters each of which is to be applied to each of the areas. The computing portion performs computation for each of the areas by using a parameter to be applied to each of the areas.

Three-dimensional shape input device

PROBLEM TO BE SOLVED: To calculate an optical flow having a small error even in an area wherein the assumption that the optical flow is constant is not satisfied. SOLUTION: This image processor 20 has: an image acquisition part 22 acquiring a plurality of image data obtained in different time; an optical flow calculation part 24 calculating the optical flow of each area inside an image by use of a gradient method on the basis of the plurality of the image data; an error calculation part 26 calculating an error of the optical flow of each the area inside the image; and an error correction part 28 substituting the optical flow of a peripheral area as the optical flow of a target area when the error of the optical flow of the target area inside the image is larger than the error of the optical flow of the peripheral area on the periphery of the target area. COPYRIGHT: (C)2007,JPO&INPIT

Image processor and image processing method

Method for acquiring shadeless target image

PROBLEM TO BE SOLVED: To provide the new flaw finding method for a rental car and the flaw finding device for implementing the method. SOLUTION: Data on the exterior surface of the rental car are stored before the rental car is rented and data on the exterior surface of the rental car are obtained again when the rental car is returned and compared with the stored data before the rental car is rented to pick up different data parts; and the parts are decided as flaws and data on the decided flaws are outputted on an output device. Further, a rental car passage gate 1 has a CCD camera 2 and is equipped with a storage device 3 for a camera image and a comparison difference detecting device 4 for stored image data and new image data.

Method for finding new flaw of rental car and flaw finding device for implementing same method

PROBLEM TO BE SOLVED: To solve the problem that when a building contour is extracted based upon an aerial photographic image and DSM, there are differences in environment of a circumference of each building and even in the circumference of one buildingSOLUTION: A position estimation section 20 determines a position estimation region including each projection part of a ground surface layer of size corresponding to a building on the basis of DSM data with respect to a rough position of a building in an object region of extraction A region division section 22 divides the object region of extraction into a plurality of small-piece regions which are regions smaller than projection regions respectively and in which properties including heights obtained from the DSM data and spectrum information obtained from an aerial photographic image are uniform An outline shape extraction section 24 extracts an outline shape of a building by integrating small-piece regions in one group having similar properties in the position estimation region A contour determination section 26 finds a building contour represented by a polygon on the basis of the outline shape

Building contour extraction device, building contour extraction method, and building contour extraction program

PROBLEM TO BE SOLVED: To provide an image recognition device for estimating a three- dimensional shape of a mobile object in a time sequential image at high speed and with high accuracy by using optical flow technology. SOLUTION: This image recognition device 1 is constituted by assimilating an optical flow by a computer processing and a coordinate of a corresponding point between images explicitly inputted by an operator. Specifically, a luminance inclination image by every pixel is generated from an object image to be included in a temporal continuous animation at a luminance inclined image-generating part 13 and the optical flow to indicate the motion of the pixel of the object image is estimated, based on the luminance inclined image and the coordinate of the corresponding point at an optical flow estimating part 14. Three-dimensional shape data restores the object image generated from the estimated optical flow and present it to the operator by a three- dimensional shape data generating part 15. When the estimated shape is improper, the coordinate of the corresponding point is corrected.

Method and device for image recognition

PROBLEM TO BE SOLVED: To provide a change extraction system between images capable of automatically extracting an area in which an object is changed without using human knowledge SOLUTION: This system is provided with an image input part 1 in which plural images whose pickup periods are different are inputted, an initial setting part 2 to match positions of the plural images, a shadow area removing part 3 to compare luminance values in pixels of the respective images, to extract the area whose value of ratio is locally large or small as a shadow area and to simultaneously correct luminance for the image with a shadow in the area in which the shadow is changed so that line same brightness as in the area without the shadow is obtained and a change extracting part 4 to calculate partial image correlation and to regard the object as changed when the correlation is low And the area in which the object is changed is automatically extracted and outputted from an output part 5

System and method for extracting change between images

PROBLEM TO BE SOLVED: To extract a building area even from an image of a relatively complicated scene photographed from the air SOLUTION: A luminance value uniformity deciding part 101 performs equal division processing of the entire area of an aerially photographed image I (i, j) into squares S (k) (the squares = S (k) is included in the image I), processing that calculates the variance value v (k) of I (i, j) in each square S (k) and processing that performs comparison operation of v (k) with threshold THv, etc In the case of v (k) =THv, a corresponding square S (k) is further divided into four parts The squares divided into four parts are also maintained The difference of an average value m (k) between squares adjacent to each other and threshold THm are subjected to comparison operation, and in the case of the difference of m (k)<THm, squares adjacent to each other are subjected to connection processing A shape characteristic deciding part 103 calculates shape characteristic quantity about an area in the image which has underwent connection processing, eg by using circularity When the circularity is close to 1, the area is of a simple shape and is recognizes as a building area

Method and device for image processing

PROBLEM TO BE SOLVED: To provide an image processor in which the specification of a variation area of a building, etc., included in a picked up image and its extraction processing is automated. SOLUTION: Plural picked up image for the same areas picked up within a period in which no change of conditions of an object like a building, etc., is regarded to exist are inputted from a storage device 4 to an image processing part 3. Edge images are generated from each of the plural picked up images and existence probability for the entire picked up image is calculated by every generated edge image by the image processing part 3. And the edge image whose existence probability is equal to or lower than a fixed value is specified as a shadow image and the image before being processed is generated by removing the shadow image from the corresponding area of the picked up image. Furthermore, the variation area of the object is specified by comparing characteristics of the preprocessed images generated for the picked up images in different points of time.

Image processor, image processing method and recording medium

PROBLEM TO BE SOLVED: To provide a change area extracting device by which a change area between image data is automatically extracted. SOLUTION: A stereo correspondence point estimating part 14 specifies a correspondence point based on a max. image correlative value which is calculated based on the image value of a specified position between image data of a stereo pair where a prescribed image conversion processing is executed by a rectification processing part 13. A change area extraction processing part 15 generates a change extraction map concerning the change area in the form of binary data to be one of the values '0' or '1' based on the max. image correlative value in the correspondence point specified by the stereo correspondence point estimating part 14 and based on a previously set correlative threshold. The generated change extraction map is held in a change area information storing part 16.

Takegawa Naoki

Papers

Method and device for image recognition

System and method for extracting change between images

Method and device for image processing

Image processor, image processing method and recording medium

Method and device for extracting change area in image data and recording medium