Hitachi Automotive Systems, Ltd.

Hitachi Astemo Ltd.

Hitachi, Ltd.

Haruki Matono

Takeshi Shima

Takeshi Nagasaki

Yuji Otsuka

Shinji Kakegawa

Toshiyuki Aoki

Hiroto Mitoma

Tetsuya Yamada

Tatsuhiko Monji

Masayuki Takemura

Hideaki Kido

Koji Doi

Takeshi Endo

Akira Oshima

Yuhi Shiina

Yoshiyuki Muto

Yoshiyuki Mutou

Keiichi Betsui

Toshiyuki Innami

Masahiro Kiyohara

Seiji Murata

Mirai Higuchi

Shinya Ohtsuji

Hiroto Morizane

Ryo Ota

Atsushi Ichige

Rikiya Kashimura

Koichi Yokoura

Naoya Tada

Hiroki Moizumi

Yasuto Arakaki

The present invention optimizes conditions of an imaging device when acquiring an image so as to improve the accuracy of image recognition. An image recognition system comprises: an image acquisition unit for acquiring an image captured by an imaging device; a recognition model generation unit having at least one environment recognition unit, recognizing at least one image by the environment recognition unit, and generating a recognition model; and a state estimation unit for estimating the state of the imaging device and the characteristic of the imaging device that corresponds to said state, on the basis of the image and the recognition model. The state estimation unit further generates an imaging correction parameter to correct the imaging condition of the imaging device, on the basis of the estimated state and characteristic and/or the recognition model.

Image recognition system

There is provided an image processing device capable of easily reducing a burden related to image recognition and obtaining stable recognition performance even when a posture of a vehicle changes from moment to moment. An image processing deviceincludes a storage unitand an image processing unit. The storage unitstores a pickup image acquired by an image pickup unitmounted on the vehicle and an inclination angle transmitted from an inclination angle sensorthat detects the inclination angle of the vehicle. The image processing unitperforms image processing on the pickup image. The image processing unitassociates the pickup image in an inclined state with the inclination angle based on a time difference Δt between a time from when the pickup image in the inclined state is acquired by the image pickup unituntil the acquired pickup image is stored in the storage unit, and a time from when the inclination angle is detected by the inclination angle sensoruntil the detected inclination angle is stored in the storage unit. The image processing unitgenerates a recognition image in a non-inclined state, which is used for image recognition of the pickup image, from the pickup image in the inclined state based on the inclination angle associated with the pickup image.

Image processing device

Provided is an object recognition device capable of accurately estimating a distance from an own vehicle to an object such as a pedestrian. There are provided: a part specification unitthat determines a specific part region in an object by analyzing an image of a region corresponding to the object detected by an overlapping region object detection unit; an image information storage unitthat stores image information on the part region determined by the part specification unit; a three-dimensional information acquisition unitthat acquires three-dimensional information of the part region specified by the part specification uniton the basis of the distance information acquired by a distance information acquisition unit; a non-overlapping region object detection unitthat detects the part region specified by the part specification unitwith reference to the image information stored in the image information storage unitfor the non-overlapping region; and a distance calculation unitthat calculates a distance to an object including the part region on the basis of the detection region information on an image of the part region detected by the non-overlapping region object detection unitand the three-dimensional information acquired by the three-dimensional information acquisition unit

Object recognition device

Erroneous detection due to erroneous parallax measurement is suppressed to accurately detect a step present on a road. An in-vehicle environment recognition deviceincludes a processing device that processes a pair of images acquired by a stereo camera unitmounted on a vehicle. The processing device includes a stereo matching unitthat measures a parallax of the pair of images and generates a parallax image, a step candidate extraction unitthat extracts a step candidate of a road on which the vehicle travels from the parallax image generated by the stereo matching unit, a line segment candidate extraction unitthat extracts a line segment candidate from the images acquired by the stereo camera unit, an analysis unitthat performs collation between the step candidate extracted by the step candidate extraction unitand the line segment candidate extracted by the line segment candidate extraction unitand analyzes validity of the step candidate based on the collation result and an inclination of the line segment candidate, and a three-dimensional object detection unitthat detects a step present on the road based on the analysis result of the analysis unit

Processing device

Provided is an image processing device that can accurately detect a target object, even when a high-distortion lens is used. According to the present invention, a cameracaptures images in accordance with a synchronization signal Sig, a cameracaptures images in accordance with a synchronization signal Sig, an area of interest setting unitsets an area of interest that represents a region to which attention is to be paid, a phase difference setting unitsets a shift .DELTA.t (a phase difference) for synchronization signal Sigand synchronization signal Sigthat synchronizes the imaging timing of cameraand camerawith respect to the area of interest in the images captured by cameraand a region of the images captured by camerathat corresponds to the area of interest, and a synchronization signal generation unitgenerates synchronization signal Sigand synchronization signal Sigon the basis of the shift .DELTA.t.

The purpose of the present invention is to obtain reliable recognition performance even in a lens which is capable of long-distance and close proximity photography. This object recognition devicecomprises an image acquisition unit which acquires an imagewhich includes a plurality of image regionswhich have differing resolutions, and a recognition processing unit which carries out a recognition process upon an object within the screen. One image region, among the plurality of image regionsof the screen, has a lower resolution than the other image regions. The recognition processing unit carries out the recognition process of the object on the basis of assessment references which differ among the one image regionand the other image regions

A video output system provided with a camera and an interface device, wherein the interface device is provided with a condition storage unit which specifies conditions under which a video is to be selected by the camera, and a video output unit which outputs a captured video received from the camera, and wherein the camera is provided with: an image sensor; a video generation unit which generates captured videos on the basis of the output of the image sensor; a condition acquisition unit which acquires said conditions from the interface device; a recognition/comparison unit which determines whether or not each captured video generated by the video generation unit meets the acquired conditions, and if it is determined that a captured video meets the conditions, generates identification information on the basis of recognition results of a subject to be recognized that is included in the captured image determined to meet the conditions, and that is associated with the conditions; and a camera calculation unit which outputs, to the interface device, at least each captured video determined by the recognition/comparison unit as meeting the conditions, from among the captured videos generated by the video generation unit, and the identification information associated with that captured video.

Video output system

An image processing device processes an image captured by at least one image-capturing device including a lens which has characteristics of including an inflection point of a change rate of an incidence angle per image height with respect to the incidence angle at a predetermined inflection point incidence angle. The image processing device includes a resolution adjustment unit that adjusts resolution of an inflection point correspondence image area corresponding to the predetermined inflection point incidence angle and resolution of at least one of an inside image area positioned more inside than the inflection point correspondence image area and an outside image area positioned outside of the inflection point correspondence image area.

In order to provide a small, stereo camera with stable errors of distance measurement, there are provided: an image acquisition unit that acquires plural images imaged by plural imaging units; a first region processing unit that divides each of the plural images acquired by the image acquisition unit into first regions composed of predetermined plural pixels, and measures a distance for each of the divided first regions to generate a first distance image; a target object extraction unit that extracts a target object from the distance image or the image, and extracts a second region composed of the plural first regions including the extracted target object; an error factor identification unit that extracts a predetermined representative value from the inside of the second region, compares the brightness value of an arbitrary pixel in the second region with that in a correspondence region up to the position moved from the arbitrary pixel by a disparity represented by the representative value, and determines an error factor on the basis of the comparison result; and a second region processing unit that determines a weight value on the basis of the error factor, and measures a distance for each second region using the weight value to generate a second distance image.

Stereo camera

An increase in the execution period of light distribution control is inhibited by reducing the number of images required for the light distribution control while achieving both object detection and light distribution control function. The present invention has an object detection unit for performing an object detection process, a light point detection unit for performing a light point detection process, and a light distribution control unit for controlling the illumination state of headlights on the basis of the result of the object detection process and the result of the light point detection process.

Headlight control device

To process sensor information positively in an information processor to which plural sensors are connected. An information processing system including plural sensors, plural buses respectively connected to the plural sensors, an arbitration unit performing arbitration of the plural buses and a priority determination unit determining the priority of each of the plural sensors, in which the arbitration unit performs arbitration of the plural buses in accordance with the determination result of the priority determination unit.

Information processing system

Provided is an image processing device capable of maintaining recognition accuracy for distance, and the like, while meeting conflicting demands such as suppressing the response delay from image input from a camera to calculation result output, enhancing recognition accuracy, reducing calculation load (enhancing processing efficiency), and enhancing response speed. The time of the starting of the processing (S) of a first distance calculation unit () for calculating the relative distance to an object on the basis of current position information determined on the basis of a left or right image and past position information for the object is set to an earlier time than the ending of the processing (S) of an object detection unit () for determining current position information for an object on the basis of parallax information obtained from the left and right images and detecting the object.

Vehicle image processing device for environment recognition

The objective of the present invention is to obtain an external environment-recognizing apparatus with which it is possible to obtain detection results having a necessary accuracy when a plurality of target objects from a captured image are detected. This external-environment-recognizing apparatus has: a plurality of processing units for processing an image; a recognition application-storing unit for storing, by object type, a plurality of recognition applications for recognizing an object by processing using at least one processing unit among the plurality of processing units; an external-information-acquiring unit for acquiring external information that includes at least external environmental information or vehicle behavior information; and a selecting unit for selecting at least one recognition application from the recognition-application-storing unit, and for selecting, on the basis of the external information, a processing unit among the plurality of processing units for processing the selected recognition application.

External-environment-recognizing apparatus

A vehicle-mounted image recognition device addresses the problem of obtaining a vehicle-mounted image recognition device capable of improving the efficiency of image processing. This vehicle-mounted image recognition devicefor solving the abovementioned problem sets a stereoscopic vision image regionin each of captured imagescaptured by a pair of right and left cameras, while setting a monocular vision image regionin the captured imagecaptured by one of the cameras, scans each of the captured images captured by the pair of right and left cameras along an image scanning line in a right-left direction, distributes the scanned image into stereoscopic vision image data obtained by scanning the inside of the stereoscopic vision image regionand monocular vision image data obtained by scanning the inside of the monocular vision image region, performs stereoscopic vision image processing using the stereoscopic vision image data, performs monocular vision image processing using the monocular vision image data, and recognizes an object corresponding to the type of an application using image processing results.

Vehicle-mounted image recognition device to set a stereoscopic-vision and monocular-vision image areas

In order to provide an imaging apparatus which minimizes erroneous recognition of a moving body and prevent a braking operation from being erroneously performed, even if the moving body is likely to cross a road, the imaging apparatus includes a correlation value calculation unit that calculates a correlation value from two images captured by two imaging units, a three-dimensional object detection unit that detects a three-dimensional object from the two images, a region dividing unit that divides an image region including the three-dimensional object into multiple regions, a relative speed calculation unit that calculates relative speed for every region of the multiple regions, and a reliability calculation unit that calculates reliability of the three-dimensional object, based on the relative speed calculated for every region.

Growing community of inventors

Tokyo, Japan

Haruki Matono