37 results about "Hypothesis verification" patented technology

A method detects a target in a sequence of radar images, wherein each image is partitioned into a grid of cells, and wherein each cell has a corresponding position in an image coordinate system associated with a location in a world coordinate system. For each most recent image in a sliding temporal window of images, intensities of each cell are determined, and the subset of the cells having highest intensities is stored as a set of current needles. A set of hypotheses, obtained by using a state transition model and corresponding maximum limits, is determined for the current set of needles and appended to a set of queues. The hypotheses for the previous sets of needles to the corresponding set of queues are updated, and a maximum likelihood in the set of queues are selected to detect the location of targets.

The present invention discloses an automobile after-sale service recommendation method and system. The system comprises a data acquisition module, an automobile basic data knowledge base building module, a pattern sequence generation module, a treelike decider building module, a user portrait building module, an after-sale service recommendation module and a physical store recommendation module, wherein the pattern sequence generation module automatically generates an extraction rule according to an automobile basic data knowledge base, and performs feature extraction and normalization on related discussed information of automobile after-sale services to enable each datum to be formalized into a pattern sequence; the user portrait building module adopts statistical information, hypothesis verification and significance analysis to obtain a plurality of distribution patterns of automobile after-sale service consumption of a user, so as to build a user portrait; the after-sale service recommendation module inputs a user vehicle feature set and the user portrait to a treelike decider, and recommends N paths with high matching degrees as service schemes to the user; and the physical store recommendation module recommends M optimal physical stores to the user for each automobile after-sale service.

A piston pose estimation method based on improved CVFH and CRH features, includes off-line processing and on-line recognition. The off-line processing firstly calibrates Kinect, then automatically renders the piston CAD model to generate multi-view point cloud, calculates its point cloud features, and finally generates off-line template library by indexing the point cloud and features. On-line identification is to estimate the position and attitude of pistons stacked randomly, firstly, the piston point cloud is collected and corrected by using the calibrated Kinect, then the point cloud is preprocessed and segmented, for the segmented point cloud, the feature of the point cloud is calculated and initially registered with the template library, then the point cloud is precisely registered byICP algorithm with the initial value of the posture, and the precise posture of the point cloud is obtained. For the generated precise posture, the mismatch is eliminated by the hypothesis verification algorithm, and the piston posture estimation and output are finally completed. The invention has the advantages of low time complexity, convenient obtaining of matching template and high accuracy of pose estimation.

The invention proposes a motion target detection method based on vision bionics, which starts from the physiologic structure of a human eye vision system and takes the space-time significant difference feature between a motion target and a background in a sequence image as an entry point. Space frequency and other space features of the sequence image are firstly extracted; and a space significant chart is formed according to the space features. Then, all the significant areas of the former and the latter frames of space significant charts are matched, thus extracting the motion speed of all the significant areas. The space features and the motion features of all the significant areas are carried out the mutual competition, coordination and fusion to form the space-time significant chart under the control by a feature library with the invariable target. After that, the invariable feature of the target in the space-time significant area is extracted. Finally, the evidence theory is adopted to carry out the sequential assumption verification of the invariable feature of the target, thus identifying the motion target. The motion target detection method improves practicality and robustness of the algorithm, which can be widely applied in various types of military and civilian systems, thus having broad market prospect and application value.

The invention discloses a substation three-dimensional live-action monitoring stationing method based on nonlinear parameter optimization calibration. The method comprises the steps that a three-dimensional substation physical model is imported; an area coverage to be monitored is acquired based on the model, and the type and the physical parameters of a camera are determined; a position hypothesis test based on the area to be monitored is carried out: three monitor points are selected on the boundary of the area to be monitored, and the coordinates of the points are calculated; based on the relationship between a model coordinate system and a camera coordinate system, the physical position and the orientation of the camera in a three-dimensional space are derived; hypothesis verification based on camera object positioning is carried out: based on a perspective projection principle, the relationship between the imaging picture of camera motion trajectory and a covered three-dimensional space entity area is analyzed; and through visualization, the three-dimensional space entity area of camera projection and the area to be monitored is compared. Compared with a traditional method, the method has the advantages that the monitor area of the same size is ensured, and at the same time the system configuration is optimized, and unnecessary device input is saved.

The invention discloses a method for detecting a nuclear explosion seismic event in a specific area. The method comprises the following steps: establishing a monitoring station network for the nuclear explosion seismic event in the specific area; obtaining real-time monitoring data of each station in the monitoring station network, and providing signal monitoring point time of the ith station; calculating time range when each hypothetical event occurs; adn when the occurrence time ranges of the hypothetical events calculated by three or more than three stations have an overlapped region, deeming that one seismic event can be formed, wherein the time range of the seismic event is the overlapped time region, and thus it is concluded that the specific area may have the nuclear explosion seismic event. The method adopts a trigger-type "hypothesis-verification" event detection mode to process the signal detection result in real time, thereby greatly reducing response time, effectively improving response speed of external monitoring tasks and achieving a near-real-time rapid detection effect of the seismic event in the specific area.

The invention provides an invariance identification method based on characteristic point and homography matching. The invention adopts a hypothesis testing method, analyzes the invariance characteristic in a scene image under the condition of knowing the line characteristic and position relation of each plane of a model, and supposes matching of some line sections so as to obtain the matching relation of certain plane and calculate the homography thereof. Then, the model plane is mapped to a scene, and hypothesis is verified by using the matching of the line section in the plane. All planes are respectively recognized. At last, the accuracy of the hypothesis is verified by the invariance between planes. Compared with the traditional hypothesis verification method, the method of the invention scatters the solving of overall situation coordinate system change in the hypothesis verification method into the solving of homography change of each plane, which lowers solving complexity and canquickly and accurately recognize objects.

A generating method of man-machine interactive quick house vector map is to alter original image after preprocess into map graph to be detected on the main curve to form a house vector map with man-machine interactive method to identify and eliminate mis-decided points inside the house by calculating border seed spots of a house to search for house border graph breaking points to be eliminated resulted in misdecided points outside the house then to connect gaps to form closed house profile, which applies house edge points assumption and assumption vertification so as to make up defect resulted in graph generating process, used in digit map generation system, and quick renew of maps.

The invention provides an online handwritten Tibetan syllable recognition method and an online handwritten Tibetan syllable recognition device, relates to the technical field of character recognition, and is used to solve the problem that Tibetan syllables input by a user in a continuously handwriting manner cannot be recognized efficiently in the prior art. The method comprises performing preprocessing on loca of points of syllables of Tibetan input by the user in the continuous handwriting manner; performing over-segmentation on the Tibetan syllables after the preprocessing successively in the horizontal direction and the vertical direction, and the obtaining sub structure block sequences of two layers of mark results; adopting a segmentation hypothesis verification method based on a semi-Markov condition random field, performing segmentation hypothesis verification on the sub structure block sequences of the two layers of the mark results, and then obtaining an optimal segmentation path and a recognition result of a part string; and, according to the optimal segmentation path and the recognition result of the part string, determining types of syllables of the handwritten Tibetan input by the user. The online handwritten Tibetan syllable recognition method and the online handwritten Tibetan syllable recognition device are suitable for the recognition of the syllables of the Tibetan input by the user in the continuously handwritten manner.

The invention in particular relates to a target recognition method based on feature geometry returns. The problems that results such as key point detection, feature matching, mismatching and the likein a conventional three-dimensional point cloud target recognition process are non-ideal, target recognition effects in complex scenes with noise, different resolution ratios and existence of barrieroverlapping and the like are poor and the like are solved. The target recognition method comprises the following steps: increasing a step of removing edge key points at a key point detection stage; eliminating a feature matching pair with ambiguity at a feature matching stage by utilizing a nearest neighbor ratio; providing a feature geometry return method at a hypothesis-generation stage, gathering accurate matches, and generating hypothesis transformation; finally, accurately verifying hypothesis transformation at a hypothesis verification stage, and accurately estimating the posture of thetarget. The method is applicable to recognition of messy three-dimensional point cloud scene target models and applied to target recognition of machine vision. The method can be widely applied to thefield of unmanned vehicle driving, robots, automated assembly, intelligent monitoring and the like.

Automated IMT system hypothesize that far wall of the common carotid artery has the highest intensity. In this current application, we verify that this hypothesis holds true for B-mode or RF-mode longitudinal ultrasound images of the carotid wall. The methodology consists of generating the composite image (arithmetic sum of images) from the database by first registering the carotid image frames with respect to a nearly straight carotid artery frame from the same database using (a) B-spline based non-rigid registration and (b) affine registration. Prior to registration, we segment the carotid artery lumen using a level set based algorithm followed by morphological image processing. The binary lumen images are registered and the transformations are applied to the original grayscale CCA images. These B-mode or RF-mode ultrasound images are then used for IMT computation using automated methods which hypothesize that far wall has the brightest intensity distribution.

Provided is a method for identifying spectrum, suitable for utilizing an electronic apparatus to identify the compounds comprised by a to-be-tested spectrum. The method for identifying spectrum comprises the following steps: executing pre-processing operation over the to-be-tested spectrum, thus to obtain the corrected spectrum; executing regression operation and hypothesis verification over the standard spectrum of multiple compounds, thus to obtain the similarity order of the compounds; selecting one by one among the compounds to add to the to-be-tested compound combination by starting from the highest one of the similarity order, thus to calculate the coefficient of rank correlation between the to-be-tested compound combination and the corrected spectrum; outputting the currently-selected to-be-tested compound combination if the coefficient of rank correlation of the currently-selected to-be-tested compound combination is smaller than that of the previously selected to-be-tested compound combination.

The invention discloses an electric vehicle automatic detection method based on vehicle-mounted vision within a blind zone. The method includes the steps that sampled data are obtained and sample characteristic data are extracted firstly, and the characteristic data are trained through a training module to generate an Adaboost cascade classifier; online detection is carried out on input image data through a mode based on hypothesis generation and hypothesis verification, the input image data are input to the trained classifier, and positioning for an electric vehicle is completed. The method aims to improve road traffic security and reduce occurrence of accidents in the blind zone. The method improves detection efficiency on the premise of guaranteeing detection accuracy.

The invention discloses an environmental target speed ambiguity resolution method and an environmental target speed ambiguity resolution system based on an automobile radar, and a storage medium, which belong to the technical field of radar speed measurement, and aim to solve the technical problems of large ambiguity resolution calculation amount and low precision at present. The environmental target speed ambiguity resolution method is characterized by comprising the steps of: transmitting two groups of pulse Doppler signals with different pulse repetition periods on the basis of an automobile radar to obtain two groups of different two-dimensional FFT data; counting a target number of different Doppler dimensions in the two groups of two-dimensional FFT data, and acquiring the Doppler dimension position with the largest target number, namely the Doppler dimension position of the environment target; and solving a real speed of the environment target by adopting a hypothesis verification method. The environmental target speed ambiguity resolution method, the environmental target speed ambiguity resolution system and the storage medium have the advantages of being small in calculation amount, high in precision, easy and convenient to operate and the like.

The invention proposes a method for evaluating the spatial distribution features of a to-be-discovered resource quantity, and the method at least comprises the following steps: 1, determining an evaluation unit; 2, subjectively setting an unit area of evaluation; 3, determining the total amount of to-be-discovered oil and gas resources and the number of oil and gas fields; 4, determining the distribution features of the total amount of to-be-discovered oil and gas resources and the number of oil and gas fields in the unit area; 5, carrying out the meshing of the evaluation unit; 6, generatingcontrol groups of the discovered oil and gas resources and the number of oil and gas fields in the unit area; 7, carrying out the judgment through a hypothesis verification method. The method can effectively solve a problem of evaluating the spatial distribution features of the to-be-discovered resources in a basin which is low in degree of prospecting or is limited in data.

A method and a system for verifying properties of a computer program is provided, pertaining to the validity of properties at program points by means of at least one forward analyser and one backward analyser. For each property, in an issuing phase an analyser issues to a centralizer module an assumption on the validity of the property at a point of the program, the centralizer module storing the assumption in a database with an attribute indicating the original analyser and a status indicating that the assumption needs to be verified; in a phase of verifying the assumption, stored in the database, a test analyser is selected to analyse the assumption in cooperation with the other analysers, the centralizer module determining analysers able to cooperate. The verification phase is iterated until all assumptions stored in the base have been analysed by at least one test analyser, a verified assumption being marked as valid.

The invention discloses a signal detection method based on error decision multi-model hypothesis test. The method comprises the following steps: 1, selecting a non-signal model set and a signal model set for signal features, and describing an original signal detection problem based on model as a binary composite hypothesis verification mode; 2, calculating the likelihood of the non-signal model set; 3, calculating the likelihood of the signal model set according to a multi-model algorithm; 4, calculating the likelihood of an intermediate signal model of the non-signal model set and the signal model set; and 5, calculating the test statistics of a 2-SPRT decision rule, and comparing the test statistics with a threshold t obtain an asymptotically optimal detection results. The method disclosed by the invention is simple in steps, is convenient to implement, solves the problems of wrong judgment and no upper limit of stop time in the existing method, can obtain a detection result with a proper average sample amount and an error probability, and can provide a new idea for the development of the detection theory in the future, is high in practicability and is convenient to promote and use.

The invention discloses a modeling method in an asynchronous D2D network based on PPP distribution and an OFDM system. Supposing that one receiving device receives discovery signals from N transmitting devices at the same time, all users are asynchronous in the network, communication links have a corresponding time deviation D and are independent of each other and the positions of the users in the region follow the Poisson point process distribution, the method comprises the following steps: dividing OFDM transmitting time misalignment into four conditions, carrying out the detailed link stage analysis, obtaining a signal power expression, simplifying the total interference power withstood by a receiving end through the further hypothesis verification, and finally proposing a first-order SINR model which is easy to process. The method accords with an actual communication environment very well, can be very conveniently used for the system-level research of the asynchronous D2D network, and overcomes a difficulty in system-level research modeling. In addition, the method also can be used for the system-level performance analysis of an ad-hoc network and a cellular network.

The invention provides a position detection apparatus, a position detection method, an information processing program and a storage medium. A hypothesis testing position detection apparatus improves robustness against abnormal detection values, and achieves both high accuracy estimation and high-speed estimation processing. An image processing apparatus (10) extracts, from calculated values (point H1 and H2) for a central position of a plane shape (I50), one of the calculated values nearer a provisional center (PP1) as a candidate for evaluation (testing) performed using a plurality of detection points (D51-Dn).

The invention relates to an NTFS file formatting parameter analysis method, terminal equipment and a storage medium. According to the method, comprehensive hypothesis verification analysis is carried out on formatted MFT records, the cluster size is determined based on the total cluster number and the physical occupation size in the MFT records, and a partition initial sector is determined based on file multi-level indexes in combination with hypothesis verification analysis. According to the method and the device, the formatting parameters required by NTFS data recovery can be accurately obtained, a good technical support can be provided for subsequent NTFS data recovery, and the recovery accuracy is improved.

The invention proposes a method for evaluating the spatial distribution features of a to-be-discovered resource quantity, and the method at least comprises the following steps: 1, determining an evaluation unit; 2, subjectively setting an unit area of evaluation; 3, determining the total amount of to-be-discovered oil and gas resources and the number of oil and gas fields; 4, determining the distribution features of the total amount of to-be-discovered oil and gas resources and the number of oil and gas fields in the unit area; 5, carrying out the meshing of the evaluation unit; 6, generatingcontrol groups of the discovered oil and gas resources and the number of oil and gas fields in the unit area; 7, carrying out the judgment through a hypothesis verification method. The method can effectively solve a problem of evaluating the spatial distribution features of the to-be-discovered resources in a basin which is low in degree of prospecting or is limited in data.

The invention proposes a motion target detection method based on vision bionics, which starts from the physiologic structure of a human eye vision system and takes the space-time significant difference feature between a motion target and a background in a sequence image as an entry point. Space frequency and other space features of the sequence image are firstly extracted; and a space significantchart is formed according to the space features. Then, all the significant areas of the former and the latter frames of space significant charts are matched, thus extracting the motion speed of all the significant areas. The space features and the motion features of all the significant areas are carried out the mutual competition, coordination and fusion to form the space-time significant chart under the control by a feature library with the invariable target. After that, the invariable feature of the target in the space-time significant area is extracted. Finally, the evidence theory is adopted to carry out the sequential assumption verification of the invariable feature of the target, thus identifying the motion target. The motion target detection method improves practicality and robustness of the algorithm, which can be widely applied in various types of military and civilian systems, thus having broad market prospect and application value.

The present invention proposes a point cloud registration optimization method based on NDT assumption verification, comprising the following steps: obtaining a reference point cloud, dividing the reference point cloud into three-dimensional grids and calculating the normal distribution of the point clouds in each grid Parameters; construct the local objective function of the point cloud in the grid to calculate the verification index δ of the NDT assumption; design a conversion function to convert δ into the weight of the three-dimensional grid in the global objective function, when the target point cloud is obtained, according to its three-dimensional The coordinates are calculated to obtain the corresponding weights of the grids to which they belong, and are substituted into the global objective function; finally, the point cloud registration result is obtained by solving the global objective function with weights. By adopting the method of the present invention, the weight of the target point cloud can be used to assist the global objective function to converge to a state where the reference point cloud and the target point cloud are more coincident, thereby improving the stability and accuracy of point cloud registration.

The invention discloses a signal detection method based on wrongly determined multi-model hypothesis testing, comprising the steps of: 1. Aiming at signal characteristics, selecting a model set without signal and a model set with signal, and describing the original model-based signal detection problem as a binary Composite hypothesis testing form; 2. Calculate the likelihood of the model set without signal; 3. Calculate the likelihood of the model set with signal according to the multi-model algorithm; 4. Calculate the middle of the model set without signal and the model set with signal Likelihood of the signal model; 5. Calculate the test statistic of the 2-SPRT decision law and compare it with the threshold to obtain the asymptotically optimal detection result. The method of the invention has simple steps and is convenient to implement, solves the problem of wrong judgment and no upper bound of stop time in the existing method, can obtain detection results with suitable average sample size and error probability, and can provide new information for the future development of detection theory The thinking, strong practicability, easy to popularize and use.

The invention discloses a method for detecting a nuclear explosion seismic event in a specific area. The method comprises the following steps: establishing a monitoring station network for the nuclear explosion seismic event in the specific area; obtaining real-time monitoring data of each station in the monitoring station network, and providing signal monitoring point time of the ith station; calculating time range when each hypothetical event occurs; adn when the occurrence time ranges of the hypothetical events calculated by three or more than three stations have an overlapped region, deeming that one seismic event can be formed, wherein the time range of the seismic event is the overlapped time region, and thus it is concluded that the specific area may have the nuclear explosion seismic event. The method adopts a trigger-type "hypothesis-verification" event detection mode to process the signal detection result in real time, thereby greatly reducing response time, effectively improving response speed of external monitoring tasks and achieving a near-real-time rapid detection effect of the seismic event in the specific area.