523 results about "Parallax" patented technology

The invention discloses a robot obstacle avoidance method and system based on millimeter wave radar sparse point cloud, and relates to the technical field of obstacle avoidance recognition. A robot obstacle avoidance system based on millimeter wave radar sparse point cloud comprises a point cloud acquisition module, a negative obstacle identification module, a weak obstacle identification module, a point cluster identification module, a risk map module, a tentative verification module and an obstacle avoidance decision module. According to the invention, suspected obstacle point clusters are extracted based on a reflection intensity threshold and a spatial proximity relation in an enhanced point cloud, a theoretical parallax model of a real static obstacle is constructed under the constraint of a robot motion trajectory, and Doppler velocity distribution of each frame is combined with a static obstacle Doppler physical law for comparison. And classifying the point clusters which do not meet the multi-view geometric consistency or Doppler physical law, and distinguishing multipath false point clusters from dynamic point clusters.

The invention discloses a space-time speckle projection three-dimensional imaging method based on multi-frame optical flow alignment. Firstly, a projector based on DLP is used for projecting a space-time speckle pattern to a measured scene, and a binocular camera synchronously collects a three-dimensional space-time speckle image. The calibration parameters of the binocular camera are used to carry out stereo correction on an acquired original speckle image, and a parallax image is generated frame by frame in combination with a coarse-to-fine single-frame speckle matching strategy. And estimating a two-dimensional inter-frame displacement field between continuous disparity maps by using an optical flow method by taking an intermediate frame disparity map as a reference, compensating motion artifacts in a space-time speckle image, and ensuring strict space-time registration of a dynamic target. And based on the speckle image after motion correction, a speckle matching strategy is expanded to a time-space domain, and high-precision multi-frame three-dimensional measurement of a complex dynamic scene is realized. The method is suitable for performing rapid and high-precision three-dimensional modeling on a moving target in an unstructured environment, and can perform accurate three-dimensional measurement on a high-speed dynamic target undergoing any translation or rotation motion.

The invention relates to a naked eye 3D binocular image acquisition and real-time processing system based on hardware synchronous triggering, and belongs to the technical field of image processing and three-dimensional display. According to the system, a hardware synchronous triggering mechanism is adopted, a time sequence synchronous control unit is used for sending a synchronous pulse signal to a binocular image sensor, and strict synchronization of left and right viewpoint image acquisition is ensured. The image signal processing unit processes collected original data, the stereo parallax correction module performs epipolar correction, and the sub-pixel interleaving module generates a composite view frame according to grating parameters of the display terminal. The system realizes high-speed data transmission through a double-buffer direct memory access DMA mechanism. The problems of visual tearing and weak stereoscopic impression caused by asynchronous binocular image acquisition in the prior art are solved, nanosecond-level synchronization precision is realized, optical crosstalk is reduced, and smoothness and comfort of stereoscopic display are ensured.

The invention discloses a star map registration method based on a remote optical image, relates to the technical field of image processing, and solves the problems that the existing method is multi-oriented to a single telescope, most of the single telescope is aligned at a star point level, and the robustness is insufficient when parallax, distortion, optical difference and background pollution exist in remote multi-station imaging. The method comprises the following steps: acquiring a remote star map and extracting a star point centroid; angular distance calculation and triangle construction; registering the centroids of the star points in the whole image; performing similarity transformation and homography estimation; and carrying out sub-pixel refinement and full image registration. According to the invention, dual registration of star point centroids and pixel coordinates can be realized for star maps shot by a plurality of telescopes in different places, and meanwhile, a high-quality registration reference can be provided for subsequent three-dimensional information acquisition of a space target and identification and positioning of the space target.

The invention provides an AI human shape recognition perception system and method based on binocular vision, and the method comprises the steps: carrying out the real-time collection through a binocular camera when a doorbell key is triggered, and carrying out the preprocessing of an original image collected in real time; performing coarse parallax estimation on the real-time image rectification to obtain a full-field coarse depth map, determining a human shape candidate region list by using the full-field coarse depth map and combining the heat source region of interest, and performing fine parallax estimation on the human shape candidate region list to obtain a fine depth patch; converting the corresponding fine depth patch into a three-dimensional point cloud set according to the pose information, and performing scale prior screening based on the corresponding three-dimensional point cloud set to obtain a plurality of human shape candidate reserved areas; and performing fusion identification according to the extracted multi-modal features to obtain a human shape identification result. According to the technical scheme provided by the invention, layered parallax estimation and three-dimensional scale prior screening can be carried out on the real-time image to realize high-reliability human shape recognition under the condition of low power consumption, so that the recognition reliability of a sensing system is improved.

The invention belongs to the technical field of trajectory measurement, and particularly relates to a three-dimensional tracking and measuring method and system for a fine high-speed object, which provides stable parallax information input through the steps of binocular video acquisition, target detection and tracking, parallax calculation and three-dimensional positioning, and three-dimensional trajectory reconstruction and measurement analysis, and provides stable parallax information in image processing. By combining high-precision target detection and a time sequence tracking algorithm, the recognition and tracking stability of a high-speed fine target is improved, the problems that target detection is difficult and parallax matching is inaccurate are solved, parallax calculation is carried out by utilizing the pixel position difference of the target in left and right images and double-target calibration parameters, and three-dimensional coordinates are quickly recovered. The processing efficiency and the real-time performance are remarkably improved, the technical defects that an existing process is complex and the response delay is large are overcome, and the method is suitable for scenes such as ball training and competitive analysis which have high requirements for high-speed measurement precision and the real-time performance.

The invention discloses a stereoscopic vision optimization method and system for a naked-eye 3D large screen, and particularly relates to the technical field of naked-eye 3D vision optimizing.The method comprises the steps that environment illumination and audience positions are sensed in real time through multi-sensor fusion, and an environment light field model is established; glare crosstalk noise is predicted based on physical simulation, and self-adaptive suppression and compensation are carried out in combination with human eye visual sensitivity and image content features; a virtual camera is dynamically generated according to the real-time positions of the eyes of the audience, and a lightweight neural radiation field renderer is used for real-time re-rendering, so that motion parallax is realized; and finally, intelligently fusing the glare compensation layer and the perspective correction layer, coding and outputting to a screen. The system correspondingly comprises an environment perception module, a glare compensation module, a perspective rendering module and a fusion coding module. The naked-eye 3D large-screen display method effectively inhibits ambient light interference, improves the quality and immersion of a stereoscopic picture under different visual angles, and is suitable for naked-eye 3D large-screen display under outdoor and complex illumination environments.

The invention discloses a digital twin modeling method and system based on binocular vision and target dynamic tracking, and the method comprises the steps: constructing a binocular camera system, completing the calibration of internal parameters, distortion and external parameters, and building a coordinate system mapping relation; synchronously collecting and correcting target images, extracting coordinates of feature points, and then removing anomalies and distinguishing multiple targets; the parallax of the feature points is calculated, three-dimensional coordinates are obtained through triangulation, and optimized target point cloud is generated; solving target space poses through a PNP algorithm, and realizing single / multi-target dynamic tracking in combination with time sequence difference; carrying out anomaly detection and smoothing processing on the pose sequence, and constructing a weighted error model closed-loop optimization feature recognition parameter; transmitting pose data to realize virtual and real synchronous visualization of the workpiece; the workpiece space relation and the collision risk are monitored, an alarm is given, and data are stored for backtracking analysis. According to the method and the system provided by the invention, high-precision dynamic measurement, attitude tracking and assembly process digitization and visualization of a complex workpiece assembly process are realized.

The invention relates to the technical field of semiconductor manufacturing, photoetching alignment and feature recognition, and particularly discloses a photoetching alignment compensation system and method based on feature targeting. An exposed pattern on the surface of a wafer is used as a unique target reference, a built-in visual imaging structure coaxial with a photoetching light path is adopted, a sub-pixel-level feature extraction and dynamic reference real-time updating mechanism is combined, the dynamic deformation of the wafer is calculated, high-frequency compensation of a workpiece table is achieved, and nanoscale overlay precision is achieved. The core improvement of the invention lies in that a light path coaxial visual integrated structure is designed to solve the problem of external imaging parallax; a feature point self-adaptive weighted matching algorithm is provided, and the deformation compensation precision is improved; and constructing a reference map real-time micro-updating mechanism, and eliminating progressive deformation drift of the wafer. The system does not need to manufacture a special physical alignment mark, continuously verify a feature matching state in the whole exposure process and quickly trigger safety protection when the matching degree is abnormal, is suitable for a photoetching alignment process for manufacturing advanced process chips of 3nm and below, has an alignment error of less than or equal to 1.1 nm under a 12-inch wafer standard test condition, reduces the dependence on an ultra-precise workpiece table, and has a good application prospect. The system has the advantages of autonomous controllability, high interference resistance and controllable cost.

The invention discloses a naked-eye 3D augmented reality interactive display system, and relates to the technical field of naked-eye 3D display and reality interaction. According to the invention, the multi-source depth data acquisition module acquires and fuses a user limb depth image, the parallax physical depth mapping module completes depth and parallax mapping and visual angle correction, the occlusion relation determination module divides regions to determine the occlusion type, and the adaptive mask layer generation module generates a dynamic smooth mask layer. The shielding effect fusion display module is used for overlapping and synthesizing the shielding layer and the multi-view virtual image and then displaying the overlapped and synthesized shielding layer and the multi-view virtual image; the problems that in naked eye 3D interaction, the occlusion relation between the limbs of the user and the virtual object is not accurate, and the vision is abrupt are effectively solved, the space consistency and the vision fluency of interaction are remarkably improved, and the real and natural naked eye 3D augmented reality interaction experience is achieved.

The invention discloses an image display method for a naked-eye 3D large screen, and particularly relates to the technical field of intelligent display, which comprises the following steps of: firstly, acquiring and tracking the positions of eyes of multiple audiences in front of the screen in real time through a depth camera network, and distributing a unique ID (Identity) for each audience to realize continuous tracking; thirdly, dividing the audiences into a plurality of visual areas based on a dynamic clustering algorithm, and calculating the rendering priority of each visual area according to the number of the audiences in each visual area and the importance of the content of the concerned screen; then, left and right eye views with correct parallax are rendered in parallel for each activated visual area, and digital pre-compensation is performed by using a pre-stored optical model so as to suppress crosstalk and uneven brightness; and finally, according to the distance and angle of the visual area of the audience, querying the human factor engineering database to obtain a comfortable parallax threshold, and dynamically compressing the extreme depth of field exceeding the threshold in the scene. According to the invention, multi-audience self-adaptive high-quality 3D display is realized, and the watching experience, the resource utilization efficiency and the visual comfort are effectively improved.

A projection unit projects pattern light on an area where a visual field of a first imaging unit and a visual field of a second imaging unit overlap each other. A measurement unit measures a distance to a surface to be measured onto which the pattern light is projected, based on a parallax between a first image obtained by the first imaging unit and a second image obtained by the second imaging unit. The pattern light is pattern light which includes a plurality of light regions with different hues and a plurality of light regions with the same hue and different levels of luminance, and in which a plurality of wide-area light regions are distributed in a predetermined pattern and a plurality of narrow-area light regions are distributed in another predetermined pattern in each of the plurality of wide-area light regions.

The invention discloses a vehicle distance measurement and calibration method based on laser radar and binocular vision fusion, relates to the technical field of measurement and calibration, and is used for solving the problem that the precision of a vehicle distance calibration result is reduced when a laser radar used for calibrating a highway vehicle distance snapshot system is under the conditions of rain, snow, wet and slippery or waterlogging road surfaces. Joint modeling is carried out on spatial distribution and echo energy characteristics of laser radar low-height point cloud, a point cloud echo coefficient and a low-height number proportion are introduced, a specular reflection state code is constructed, a self-calibration mechanism is adaptively triggered accordingly, and a specular reflection state is obtained. The discrete characteristic of a laser radar vehicle distance time sequence and the degradation trend of binocular vision parallax connectivity are subjected to quantitative analysis, the dominant source of ranging abnormity is judged through symbolic analysis, and an offset calibration or fusion correction strategy is selected according to the dominant source. Reliability and engineering adaptability of vehicle distance measurement under the wet and slippery road surface condition can be improved under the condition of not depending on an additional sensor.

For see-through AR displays, the current state of art uses geometric correction to lower the display location error, however this is far from an ideal solution and requires knowing range or pose of the devices. A dual-function optic for eye display & camera is disclosed, wherein a sensor camera imaging optics pipeline with a virtual projection point at an eye's retina reduces parallax errors to zero.

The invention discloses a multi-modal three-dimensional image quality evaluation method based on consistency-complementarity characteristics, and belongs to the technical field of three-dimensional image quality evaluation and brain-computer intelligence. The method comprises the following steps: S1, acquiring a plurality of groups of distorted stereo images and electroencephalogram signals corresponding to the distorted stereo images, preprocessing the distorted stereo images, and endowing a quality grade label to a sample according to a subjective experiment result; s2, inputting the preprocessed three-dimensional image and the electroencephalogram signal into a parallax perception image encoder and a multi-scale attention electroencephalogram encoder to obtain binocular image features and electroencephalogram spatial-temporal features; s3, inputting the binocular image features and the electroencephalogram spatial-temporal features into a complementary learning module, and extracting image complementary features, electroencephalogram complementary features and cross-modal consistency features; and S4, inputting the consistency characteristics into a consistency learning module for cross-modal alignment to obtain an optimized consistency characteristic representation, fusing the optimized consistency characteristic representation with the complementary characteristics, inputting the fused consistency characteristic representation into a classification module, and outputting a stereo image quality grade.

The invention discloses a binocular vision naked eye 3D image processing method, device and system, and particularly relates to the technical field of naked eye 3D. According to the method, the foreground and background areas are defined, a parameter extraction and merit figure quantification calculation system covering parallax, scenes and splicing dimensions is constructed, and the multi-scene preset threshold and the weight factor are combined, so that comprehensive and accurate evaluation of the naked eye 3D image quality is realized, the problem of one-sided evaluation of a traditional single index is solved, and the evaluation efficiency is improved. And the core crux influencing the three-dimensional effect is accurately positioned.

The invention relates to the technical field of remote sensing surveying and mapping, in particular to a three-dimensional terrain reconstruction method and system based on a remote sensing image, and the method comprises the following steps: collecting visible light and near-infrared three-dimensional multispectral data, carrying out pixel-level registration based on geometric calibration, constructing a geometric registration multispectral image matrix, generating waveband difference, and extracting a structure edge; the method comprises the following steps of: constructing a spectral signal diffusion impedance field, forming a terrain self-adaptive matching support domain, calculating a multispectral matching parallax error, and resolving a three-dimensional terrain reconstruction model based on space forward intersection, in the method, the terrain boundary recognition capability is enhanced through multispectral geometric registration and waveband difference analysis, and stable edge features are extracted in combination with gradient amplitudes; the matching process adapts to topographic fluctuation changes, signal diffusion impedance constructed based on spectral difference restrains a matching support domain form, cross-ground object interference is suppressed, parallax calculation stability is improved, and three-dimensional elevation and structure expression reliability under complex earth surface conditions is enhanced.

The invention relates to the field of optical imaging and computer vision, discloses a large-view-field binocular image combination dynamic parameter adjusting system, and aims to solve the problems that a traditional head-up display is limited in view field and monochromatic display and information response are lagged. The system comprises a left and right eye image acquisition and preprocessing unit, a target positioning and trajectory prediction unit, a binocular parallax compensation and registration unit, a dynamic fusion display control unit, a color enhancement and information superposition unit, an eye movement tracking feedback unit and a parameter adaptive adjustment unit. Through binocular wide-angle acquisition, three-dimensional target prediction, sub-pixel-level image registration, panoramic stitching fusion, semantic color coding and dynamic parameter adjustment of eye movement driving, 120-degree * 60-degree distortionless panoramic display, multi-source information intelligent layered superposition and visual focus area adaptive enhancement are realized. According to the method, the air combat situation sensing efficiency and the weapon response speed are improved, and the method has high real-time performance (delay of lt, 80ms), strong environmental adaptability and engineering integration.

The invention discloses a large-parallax image splicing method and device, equipment and a storage medium, and belongs to the technical field of image processing. The method comprises the following steps: performing feature point matching on a to-be-spliced feature point set and a reference feature point set to obtain a feature matching point pair set; according to the parallax value of each feature matching point pair in the feature matching point pair set, layering the parallax of an overlapping region intersected with the reference image in the to-be-spliced image to obtain a plurality of parallax layers and a matching feature point pair set corresponding to each parallax layer; according to the matching feature points to be spliced in the matching feature point pair set of each parallax layer, determining each geometric feature region by using a region growing method; and according to the global homography matrix and the local homography matrix of each geometric feature region, performing fusion processing on the to-be-spliced image and the reference image to obtain a target spliced image. By using the technical scheme provided by the invention, the accuracy of large parallax scene image splicing can be improved.

The invention relates to an environment sensing method based on target detection and stereo matching. According to the environment perception method based on target detection and stereo matching, targets in an environment image can be recognized through model training, and category information and position information of each target can be recognized, so that a target detection model can recognize the category information and the position information of the targets in the environment image in real time; through a local matching strategy of a target area delineated by focusing on the position information of the target, huge calculation overhead caused by full-image matching is avoided; pixel-level parallax information in the two-dimensional image is accurately converted into a physical depth distance in a three-dimensional space in combination with calibration parameters of a binocular camera in an environment sensing system; and finally outputting an environment perception image capable of displaying a detailed perception result of each detected target through information fusion. According to the environment sensing method, the efficiency and robustness of the whole sensing process are remarkably improved while the accuracy and integrity of environment sensing are ensured.

The invention provides a curved surface laser engraving method and related equipment based on a camera module and active light spot projection, and the method comprises the following steps: controlling an active light spot projection module to project an active light spot on the surface of a workpiece placed on a processing table top of a laser engraving machine, and synchronously controlling the camera module to shoot, obtaining a three-dimensional image pair of the workpiece; performing parallax matching processing according to the three-dimensional image pair of the workpiece to obtain high-density point cloud data of the surface of the workpiece; constructing a workpiece curved surface three-dimensional model according to the high-density point cloud data; mapping the two-dimensional design pattern to a workpiece curved surface three-dimensional model to generate a laser engraving track containing curved surface height compensation; and controlling the laser engraving machine to perform curved surface laser engraving on the workpiece according to the laser engraving track. The curved surface laser engraving precision and the overall efficiency can be improved.

The invention discloses a power transmission line tree obstacle distance measurement method based on binocular vision and deep learning fusion, and relates to the technical field of electric power, and the method comprises the steps: collecting a binocular image, carrying out the calibration and stereo correction, and carrying out the single-scale Retinex enhancement processing of the corrected image; and performing power transmission line and tree obstacle region segmentation on the enhanced image based on a DeepLabv3 + model in combination with a coordinate attention mechanism, and performing semantic-guided parallax estimation and optimization on the enhanced image based on a segmentation result to obtain an optimized parallax map. According to the method, a multi-level feature map is extracted through Res2Net-50, target area positioning is enhanced in combination with a coordinate attention mechanism, an ASPP module is used for fusing multi-scale context information, the segmentation precision of a power transmission line and a tree barrier is improved, a Kalman filter for adaptively adjusting observation noise variance is used for processing multi-frame distance measurement data, and the accuracy of distance measurement is improved. And the stability and the reliability of distance measurement are improved.

The invention provides a device for realizing a virtual image naked eye 3D image and a generation method, the device comprises an image generation unit and a virtual image system, the image generation unit generates left and right frame parallax images; and the virtual image system is of a surface type structure of a reflecting mirror group or a curved reflecting mirror group with a reflecting function, receives the left and right frame parallax images generated by the image generation unit, and puts the left and right frame parallax images into the left and right eyes respectively. According to the invention, no resolution loss exists in the switching process of the 2D virtual image and the 3D virtual image, and distortion and imaging quality of the original virtual image system are not affected by joint debugging of the virtual image system and the image generation unit.

The application relates to the technical field of intelligent analysis of power transmission lines, and particularly discloses a multi-objective cooperative video online monitoring and intelligent analysis system for power transmission lines. The system comprises: a buffer construction module for constructing an independent image buffer; a phase alignment module for extracting an optical intensity evolution waveform to obtain a time delay amount, and resampling a video to generate a feature point sequence; a flashover positioning module for extracting a synchronous highlight two-dimensional coordinate and solving a physical space coordinate; a coverage analysis module for constructing an expansion bounding volume, performing spatial analysis on the expansion bounding volume and a cylindrical envelope obtained by geometric dimension reduction of a power transmission line model, and obtaining an erosion coverage area; and a directional cooperation module for driving a pan-tilt camera to perform high-resolution rephotographing based on the solved physical coordinate and the erosion area. The system is beneficial to reducing multi-camera asynchronous and parallax problems, and realizing three-dimensional positioning of arc transient discharge of the power transmission line and quantification of a damaged range of equipment.

The invention provides a total station automatic focusing and reflector plate identification method and system based on intelligent image.The method relates to the technical field of engineering surveying, and comprises the steps that a stepping motor is controlled to drive a focusing lens barrel to move step by step in the direction of an optical axis, a coaxial camera collects images at all positions and calculates definition evaluation values to complete coarse adjustment and fine adjustment focusing, and the focusing precision is improved; and automatic focus locking is realized. The center coordinates of the reflector plate are determined through image preprocessing and reflector plate area recognition, the horizontal and vertical angle deviation can be calculated, sighting is achieved by driving a horizontal servo motor and a vertical servo motor to rotate automatically, and closed-loop control can be completed through iterative correction. The system is composed of a coaxial camera, a focusing execution mechanism, a master control module, an image processing module, an angle calculation module and a servo driving module, a coaxial optical path structure is adopted to eliminate parallax errors, the system has the high-precision automatic focusing, intelligent recognition and rapid automatic sighting capabilities, and the automation level and the measurement precision of the total station can be remarkably improved.

The invention discloses a video coding method and device and a video decoding method and device based on residual prediction, and is applied to the technical field of video processing. The coding method comprises the following steps: determining a motion vector and a disparity vector of a current coding block; determining a first parallax compensation block corresponding to the current coding block in the reference view according to the parallax vector; determining a prediction residual according to the current coding block, the first parallax compensation block, and the motion vector and the parallax vector of the first parallax compensation block; or determining a first time domain motion compensation block corresponding to the current coding block in the previous frame and / or the next frame of image of the current coding block according to the motion vector; determining a prediction residual according to the current coding block, the first time domain motion compensation block, and the motion vector and the disparity vector of the current coding block; and coding the current coding block according to the prediction residual error. According to the method, the accuracy of the obtained prediction residual error can be improved, so that the coding compression efficiency and the coding effect of the multi-view video are improved.

The invention discloses a power transmission line conductor diameter measurement method and system based on an unmanned aerial vehicle image, and the method comprises the steps: obtaining a multi-view image of a power transmission conductor through an unmanned aerial vehicle, and recording the shooting pose and distance information at the same time; carrying out distortion removal and geometric correction processing on the multi-view image to obtain the pixel width and parallax of the lead, and calculating a first estimated value of the lead diameter and the uncertainty of the first estimated value according to the parallax; obtaining a reference image containing the wire and the insulator through the unmanned aerial vehicle, identifying the model of the insulator and obtaining the real size of the insulator, and calculating a second estimated value of the wire diameter and the uncertainty of the second estimated value through the real size of the insulator according to the proportional relation; and carrying out consistency check on the first estimation value and the second estimation value, and obtaining a fusion line diameter result by adopting a weighted fusion algorithm. According to the invention, through fusion of three-dimensional measurement and reference object inference, high reliability and accuracy of wire diameter measurement are still maintained in a complex environment, and an efficient and safe solution is provided for power transmission line inspection.

The invention relates to a mobile robot motion estimation method based on pose decoupling. The method comprises the following steps: multi-sensor synchronization and decoupling state modeling; performing attitude-position decoupling pre-integration based on an IMU (Inertial Measurement Unit); visual observation modeling and reprojection constraint construction are carried out; carrying out pose optimization solution based on the decoupling factor graph; pose recovery, motion track output and local re-initialization are carried out; compared with the prior art, the method has the advantages that independent modeling and decoupling estimation are carried out on the pose rotation component and the translation component, the propagation effect of rotation errors to translation solution in traditional visual motion estimation is effectively inhibited, and the stability of translation estimation is remarkably improved. Under visual degradation scenes of weak texture, low parallax, illumination variation, motion blur and the like, estimation drift can be reduced, track continuity and consistency can be improved, and robustness of the system to abnormal matching and noise disturbance can be enhanced; besides, the decoupling mechanism provided by the invention has good universality, can be integrated with various types of visual SLAM or visual-inertial fusion frames, and provides more reliable motion estimation capability for stable autonomous navigation of the mobile robot in a complex environment.

To provide an electronic apparatus, a program, and an information processing method capable of accurately measuring a position of a pedestrian.SOLUTION: An electronic device (10) includes a disparity data processing unit (11) that generates disparity data based on image data acquired from a stereo camera, a skeleton detection unit (12) that detects a skeleton of a person included in the disparity data, a center-of-gravity recognition unit (13) that recognizes a center of gravity of the person based on the skeleton of the person, an ROI setting unit (14) that sets an ROI including the center of gravity of the person for the disparity data, and a conversion unit (15) that converts a position of the center of gravity of the person into a position in a sensor coordinate system based on the disparity data in the ROI.SELECTED DRAWING: Figure 1

To provide a technique for estimating accuracy of dimension measurement performed by using a plurality of parallax images having parallax with each other photographed under a specific photographing condition.SOLUTION: An accuracy estimation device includes acquisition means for acquiring photographing conditions of a plurality of parallax images mutually having parallax, and estimation means for estimating accuracy of dimension measurement performed using the plurality of parallax images on the basis of the photographing conditions.SELECTED DRAWING: Figure 4