73 results about "Image moment" patented technology

What is disclosed is a novel system and method for encoding / decoding data in a cover contone image via halftone dot orientation modulation. Arrays of halftone threshold values are used to determine a desired orientation, e.g. 0 / 90°±45° for a given single data value of the original message to be embedded. Message data is embedded as a function of halftone dot orientation. Detection modeling of the print-scan process enables the determination of dot orientation from the image scan via statistically motivated image moments. A probabilistic model of the print-scan channel conditions received moments on input orientation. Density values of the received moments are used to determine dot orientation for each halftone cell. The embedded data is retrieved based on the determined orientations. The present method is applicable to areas of data embedding, document security, and the like.

What is disclosed is a novel system and method for encoding / decoding data in a cover contone image via halftone dot orientation modulation. Arrays of halftone threshold values are used to determine a desired orientation, e.g. 0 / 90°+ / −45° for a given single data value of the original message to be embedded. Message data is embedded as a function of halftone dot orientation. Detection modeling of the print-scan process enables the determination of dot orientation from the image scan via statistically motivated image moments. A probabilistic model of the print-scan channel conditions received moments on input orientation. Density values of the received moments are used to determine dot orientation for each halftone cell. The embedded data is retrieved based on the determined orientations. The present method is applicable to areas of data embedding, document security, and the like.

A method for compressing a bi-level document image containing text is disclosed. The document image is segmented into symbol images each representing a letter, numeral, etc. in the document. The symbol images are classified into a plurality of classes, each class being associated with a template image and a class index. Classification is done by comparing each symbol to be classified with template of existing classes, using a number of image features including zoning profiles, side profiles, topology statistics, and low-order image moments. These image features are compared using a tolerance based method to determine whether the symbol matches the template. After classification, certain classes that have few symbols classified into them may be merged with other classes. In addition, the template images of the classes are down-sampled, where the final sizes of the template images are dependent on the likelihood of confusion of the template with other templates.

The present invention provides a method and an apparatus of determining video image jitter and camera device jitter. The video image jitter determining method comprises the steps of acquiring multiple feature points in a video image at a first moment; tracking the multiple feature points, and determining at a second moment whether content in the image moves drastically according to tracked feature points; if the content in the image does not move drastically, determining a quantity of unmatched feature points in the tracked feature points; and determining, according to the quantity of the unmatched feature points, whether the video image at the second moment jitters compared with the video image at the first moment.

What is disclosed is a novel system and method for encoding / decoding data in a cover contone image via halftone dot orientation modulation. Arrays of halftone threshold values are used to determine a desired orientation, e.g. 0 / 90°+ / −45° for a given single data value of the original message to be embedded. Message data is embedded as a function of halftone dot orientation. Detection modeling of the print-scan process enables the determination of dot orientation from the image scan via statistically motivated image moments. A probabilistic model of the print-scan channel conditions received moments on input orientation. Density values of the received moments are used to determine dot orientation for each halftone cell. The embedded data is retrieved based on the determined orientations. The present method is applicable to areas of data embedding, document security, and the like.

The invention relates to an SAR satellite simulated on-orbit load data obtaining system and obtaining method. According to SAR satellite imaging parameters, SAR echo data is manufactured and is stored in a task planning computer; a navigation signal simulator and an attitude and orbit control ground equipment synchronically output on-orbit simulated flight relevant excitation to an SAR satellite to execute the simulated on-orbit flight; an imaging planning instruction is sent to the SAR satellite; an imaging mode, radar work parameter information, imaging moment, imaging duration, imaging process attitude information and load data downloading moment are transmitted to the SAR satellite. When the preset SAR satellite imaging moment is reached, an echo signal is sent to the SAR satellite; the SAR satellite packs formed load data; when the preset SAR satellite load data downloading moment is reached, load data receiving equipment receives a load data signal sent by the SAR satellite; demodulation receiving and restoration are performed, so that the load data receiving is completed.

The invention discloses a rotation invariant submarine topography two-dimensional matching aided navigation method which comprises the following steps: 1) a real-time stripe sweeping survey depth value matrix array is obtained by a multi-beam sounding system, height array values are converted into gray values to obtain a template grayscale map; topography two-dimensional height array values in a region in the main navigation system error range are read, and then are converted into gray values to obtain a background grayscale map; 2) a rotation invariant image moment function is used as a characteristic feature, a region in the background grayscale map, matched with the characteristic feature of the template grayscale map can be searched, and the position of the multi-beam sounding system in real-time sweeping survey can be obtained. Compared with matching algorithms (such as TERCOM matching algorithm and SITAN matching algorithm) in the prior art which are performed by topography one-dimensional matching, the rotation invariant submarine topography two-dimensional matching aided navigation method overcomes the defects that the one-dimensional matching is low in accuracy and the restricted by topography, increases the richness of topography features description, and has higher accuracy and adaptability.

A system and method for white spot Mura defects on a display. The system is configured to pre-process an input images to generate a plurality of image patches. A feature vector is then extracted for each of the plurality of image patches. The feature vector includes at least one image moment feature and at least one texture feature. A machine learning classifier then determines the presence of a defect in each patch using the feature vector.

The invention provides a common-orbit multi-point target imaging task planning method for an agile satellite. The method comprises the steps that S1, determining n point targets needing to be imaged and imaging moments corresponding to the n point targets; S2, obtaining an imaging moment set x and a set p of n point targets corresponding to the imaging moment set x according to the sequence of theimaging moments corresponding to the n point targets; S3, sequentially taking i from 1 to n-1;selecting j from 2 to n, i is smaller than j, and determining a set a of public vertex values corresponding to a set p; S4, obtaining an attitude constraint undirected graph structure according to the public vertex value set a; Step S5, performing block orientation on the attitude constraint undirected pattern structure to obtain an attitude constraint directed pattern structure, and obtaining a plurality of groups of imaging sequences according to the attitude constraint directed pattern structure;and S6, selecting at least one group of image sequence from the plurality of groups of image sequences as an optimal imaging sequence.

The invention discloses a discrete cosine phase information-based multi-spot detection and locating method. The method comprises the steps of performing binary cutting and edge detection by traversing a threshold maximum variance by adopting multi-target regional augmentation of discrete cosine phase information, and performing multi-spot centroid calculation by adopting an image moment method to finally obtain edges and centroids of all spots. According to the method, phase information of discrete data is fully considered for realizing image target detection. The method is high in calculation speed, effective and reliable in effect and suitable for related fields of laser measurement, robot vision locating, optical precision measurement and the like.

A plurality of gray-scale values is extracted from image signals corresponding to a dimming area to calculate a mean value of the gray-scale values, and at least one of a variance, a standard deviation, a kurtosis, a skewness, a central moment, and an image moment is calculated using the mean value. Then, a representative gray-scale value corresponding to the dimming area is determined using the calculated values, and a dimming function for the light sources included in the dimming area is determined based on the representative gray-scale value. Then, the light sources included in the dimming area are driven based on the dimming function.

The present invention provides a magnetic resonance temperature imaging method and system. The method comprises the steps of determining a compound excitation pulse including a plurality of single excitation pulses on basis of a highly accelerated parallel imaging aliasing control method; at a lot of temperature imaging moments, respectively using the compound excitation pulse to simultaneously exciting multilayer imaging areas, and respectively acquiring aliasing signal data of the multilayer imaging areas. calculating and obtaining magnetic resonance image data of each layer of the imaging areas at the temperature imaging moments by a parallel imaging algorithm according to the aliasing signal data; and calculating and obtaining temperature image data of each layer of the image areas at the temperature imaging moments by a temperature algorithm according to the magnetic resonance image data of each layer of the imaging areas at the temperature imaging moments. A lot of temperature imaging moments include an initial temperature moment before temperature changes, and the temperature image data include temperature difference between the temperature imaging moment after the temperature changes and the initial temperature moment. According to the magnetic resonance temperature imaging method and system, temperature imaging efficiency can be improved.

The present invention relates to a crane intelligent lifting system based on limb motion recognition. Based on a human-computer interaction system of six control units, the crane intelligent lifting system is characterized in that: 1) a large zoom network camera continuously collects multiple sets of standard limb motions of the command personnel, 2) a large zoom network camera collects limb motions of the command personnel in real time through the cloud platform on the crane cockpit, and then enters a contour extraction unit and a feature extraction unit to obtain a gesture Hu image moment feature vector of the command personnel; and 3) query is performed through a sample classifier by using the SVM machine learning algorithm, and the most matched limb motion sample class is returned, sothat the corresponding crane command signal is transmitted to the crane controller to complete the crane lifting work. According to the system disclosed by the present invention, the shortcomings of the above-mentioned existing limb motion recognition technology is overcome, that the signal human judgment and command mode is not friendly enough is overcome, and referring to the national standard GB5082-85 of the lifting command signal, the command personnel can be tracked in real time, and the system has the advantages of a low cost, strong robustness and high real-time performance.

A plurality of gray-scale values is extracted from image signals corresponding to a dimming area to calculate a mean value of the gray-scale values, and at least one of a variance, a standard deviation, a kurtosis, a skewness, a central moment, and an image moment is calculated using the mean value. Then, a representative gray-scale value corresponding to the dimming area is determined using the calculated values, and a dimming function for the light sources included in the dimming area is determined based on the representative gray-scale value. Then, the light sources included in the dimming area are driven based on the dimming function.

The present invention provides an image processing method for removing the Newton's halo. The method comprises: carrying out double format conversion on a to-be-processed image; and preprocessing thesize of the image after being subjected to the double format conversion, acquiring image data, extracting an effective area of the image sample by using background modeling, carrying out analysis andmachine learning on the halo sample, carrying out HSV decomposition on the noise image, and extracting the luminance component. According to the method provided by the present invention, the denoisingprocessing method is not limited, the background modeling is to establish a reasonable and realistic background for the image, the image sample screening is the primary screening and artificial refinement screening by using the color and size of the image, image region cognition is to use image color statistical information and image moment information as the classification features of the image,and the image suppression algorithm is used to suppress the halation of the region; and the halo target appearing in the video is recognized and detected after acquiring the video, and halation suppression is performed on the detection area, so that the readability of the video monitoring picture is improved.

The invention discloses a multifunctional monitoring method for self-adaptive adjustment of a wire cut electrical discharge machining process, comprising the following steps: establishing a multifunctional monitoring device, starting electrical discharge machining, and respectively obtaining videos of a reference surface discharge condition and an electrode wire motion state through a special high-speed camera shooting observation system; performing OpenCV image processing on the videos of the discharge condition, carrying out statistical analysis on spark center points through a centroid method to obtain a discharge point distribution statistical graph; carrying out electrode wire position point analysis on the videos of the electrode wire motion state to obtain the electrode wire motionstate; obtaining discharge waveforms through a voltage high-speed acquisition card system, and classifying the discharge waveforms through an image moment, fractal dimension, a wavelet transformationmethod and a support vector machine classification method to obtain the percentage of each type of discharge waveforms. The method is diversified in function, high in detection precision and visual inmonitoring, can be used for guiding and selecting the optimal machining method and machining parameters, and provides a basis for self-adaptive adjustment of wire cut electrical discharge machining.

The invention relates to a chip detection method and device, computer equipment and a storage medium. The method comprises the steps of obtaining a chip image of a chip to be detected; Generating a corresponding image moment according to the chip image; Generating an identification operator corresponding to the chip image according to the image moment; If the identification operator of the chip image is matched with a preset template identification operator, generating a detection qualification result; And if the identification operator of the chip image is not matched with the preset templateidentification operator, generating a detection unqualified result. According to the invention, the accuracy of chip qualification detection can be improved.

The invention discloses a (D)KDP crystal body damage performance high-precision measurement device and measurement method, relating to the technical field of (D)KDP crystal body damage measurement. According to the method, crystal body damage point base data after action of high power nanosecond laser pulse is acquired through chromatography, problems like damage point repeated statistics, background light elimination and binarization are solved, center of mass of each scattering point is resolved by using an image moment algorithm, then body damage point three-dimensional distribution is acquired through a reconstruction algorithm, then three body damage representation parameters, including crystal body damage density ppd, body damage point geometric size distribution pps and crystal bodydamage point three-dimensional distribution can be acquired highly precisely. The method has the advantages of high measurement accuracy and more complete crystal body damage representation.

The present invention discloses a limited angle cone-beam CT image reconstruction method based on a geometric image moment. The method comprises the steps of: converting obtained cone-beam projection data into known Radon data; obtaining projection-geometric moment transformation of the known Radon data; according to the projection-geometric moment transformation of the known Radon data, calculating the geometric image moment; according to the geometric image moment, estimating projection-geometric moment transformation of unknown Radon data; determining the unknown Radon data through inverse transformation; performing data combination on the known Radon data and the unknown Radon data; obtaining complemented three-dimensional data; and reconstructing a CT image. According to the method of the present invention, a cone-beam CT image that meets the requirements of clinical diagnosis and has high quality can be reconstructed under the conditions of reducing a scanning range.

The invention relates to a border tracking and chain code generation arithmetic based on an automatic machine and belongs to the field of digital image treatment. The border tracking arithmetic is important in image science and chain encoding is widely applied on image treatment and analysis. The arithmetic can scale a multi-region and multi-connection image with high efficiency and convenience to obtain a chain code. The technical proposal includes determining the type of the automatic machine; scanning the whole area of a binary image from up to down and from left to right; when a border starting point is detected, starting the automatic machine to run along the border for one circle and generating correspondent chain code; returning to the starting point with the present direction the same as the initiating direction to finish scaling of the border; and stopping scanning to finish scaling. The invention is suitable for representing image, storing binary image, calculating image moment, calculating geometrical characteristics of letters and image, analyzing document image, and automatically identifying letters in digital image treatment.

The embodiment of the invention provides a light and shadow rendering method of a virtual object in a panoramic video, a device and an electronic device. The method comprises the following steps: judging whether a preset virtual object characteristic point is included in a current picture of a panoramic video on a display screen; and determining whether the preset virtual object characteristic point is included in the current picture of the panoramic video. If the virtual object feature points are included, decomposing the current picture into a first preset number of sub-pictures; Determininga light intensity weighting center according to an image moment of the sub-picture, and determining light source position information of the current picture based on the light intensity weighting center and the geometric center of the sub-picture; Determining contour information and position information of the virtual object based on the feature points of the virtual object; A light and shadow effect of that virtual object is generate according to the light source position information and the outline information and the position information of the virtual object. Whether a virtual object is included in a currently played panoramic video can be determined in real time, and a light shadow effect of the virtual object is generated according to a light source position in the video.

The embodiment of the invention provides an optimization method and device of illumination information, and electronic equipment. The method comprises the following steps: acquiring a target picture,decomposing the target picture into a first preset amount of sub-pictures; acquiring a geometric center of each sub-picture, and determining a light intensity weighting center according to an image moment center of the sub-picture, and determining the illumination angle of the sub-picture according to the light intensity weighting center and the geometric center of the sub-picture; performing grayprocessing on the sub-picture, and obtaining the illumination intensity of the sub-picture according to the gray value of each point on the sub-picture; acquiring the illumination intensity weight value of the sub-picture, and obtaining the illumination intensity of the target picture according to the illumination intensity of the sub-picture and the illumination intensity weight value of the sub-picture; and obtaining the illumination angle of the target picture by computing the illumination angle of the sub-picture.

The invention discloses a high-precision image moment positioning method, which comprises the following steps: (1) separating an image of a target to be detected from background; (2) calculating the initial moment of the target; (3) selecting a weighting function suitable for target image distribution as a weighting coefficient, wherein the weighting function needs to exhibit four characteristics of normalization, centrosymmetry, attenuating property and incoherence; (4) according to the weighting coefficient, recalculating a new moment of the target image, outputting the new moment as a target centroid if an error between the new moment and an original moment satisfies precision requirements; otherwise, taking the new moment as an iteration starting point to reselect the weighting coefficient and recalculate the new moment until the error between the new moment and the original moment satisfies the precise requirements; and outputting the new moment as a target parameter. The method has the advantages of being high in target positioning precision, high in anti-noise capability, simple in calculation and easy for engineering realization, does not aim at a specific target image and is especially suitable for occasions which adopt the image with a low signal to noise ratio and complex background to carry out target positioning.

The invention discloses a four-rotor unmanned aerial vehicle visual servo control method using a disturbance observer and a nonlinear speed observer. The method comprises the steps: establishing a corresponding dynamic model according to the space motion condition of a four-rotor unmanned aerial vehicle; selecting an image moment as a visual servo feature, fusing a virtual feature plane, and establishing image feature dynamics under a virtual camera plane; designing a dynamic visual servo controller based on an image, firstly, analyzing and designing a disturbance observer according to the acceptance condition of the quad-rotor unmanned aerial vehicle, then, designing a speed observer in a virtual camera plane, and finally, designing a sliding mode controller based on exponential approaching by using a backstepping method. Therefore, the robustness of the quad-rotor unmanned aerial vehicle to external disturbance can be improved, the non-measurable image feature linear velocity in the virtual image plane can be accurately estimated, and the robustness of the system can be further improved by the sliding mode controller based on exponential approaching finally.

Provided are an image region stray light eliminating device based on moment matching and a method thereof. The device acquires a reference mean value and a reference variance calculated through image moment matching according to the statistic law of the image grey information and the grey change trend of the whole image and conducts standard moment matching. By means of the method, effects of stray light can be well eliminated aiming at even surface feature remote sensing images affected by the stray light in the image inside area, and even transition of the image in the grey list direction is maintained. By means of the method, the image can be restored effectively, the image quality is improved, and the algorithm efficiency is high.

The invention provides an in-orbit calibration method for high precision imaging moment. The method comprises the following steps: step 1, acquiring the imaging moment sent from the ground on satellite; step 2, based on the acquired imaging moment, using the forecast satellite information from GPS receiver to acquire the satellite position and speed of the moment; step 3, based on the position and speed coordinated with the position of the imaging target point, calculating the calibrating the time deviation; step 4, calculating to obtain the calibrated imaging moment; step 5, judging whether the time deviation calibration meets the precision requirement. If not met, selecting the time acquired in step 4 to be the imaging moment, repeating the procedure from step 2 to step 5; if the requirement is met, the calibration calculation of imaging moment is completed, using the calibrated imaging moment to start the imaging device. The method has the advantages of low calculation burden and easy management for the satellite-borne computer. The method can be used for the determination of the high precision imaging moment during the normal working period of satellites, and provides a high precision time benchmark for the imaging.

The invention relates to a template matching based pointer type instrument multi-condition automatic reading method. The method comprises the following steps: step 1, carrying out dial plate posture correction based on angular point matching and perspective transformation; 2, positioning the center of a pointer based on template matching of a two-step method; and step 3, carrying out pointer rotation angle estimation based on the Zernike image moment.

The invention discloses a wavefront correction method based on image moment characteristics. The method uses a device, and the device comprises an image acquisition module, an image processing module,a neural network module and an execution mechanism. The specific method comprises the steps that incident light is divided into two light spots through a spectroscope, the incident light are converged into a focused light spot and a defocused light spot respectively, and an image collection module is responsible for collecting the light spots; the image processing module is responsible for carrying out image moment operation on the light spots and taking characteristics extracted after the image moment operation as the input of the neural network; after training of a large number of image features, the neural network module fits a corresponding Zernike coefficient according to the image features transmitted by the image processing module; and finally, the execution mechanism corrects thewavefront distortion generated by the atmospheric turbulence according to the fitted Zernike coefficient.

The embodiment of the invention provides a method and a device for determining a shadow effect, and an electronic device. The method comprises the following steps: obtaining a target picture, and decomposing the target picture into a first preset number of sub-pictures; determining a light intensity weighting center according to an image moment of the sub-picture, and determining light source position information of the target picture based on the light intensity weighting center and the geometric center of the sub-picture; generating a shadow effect of the target object according to the lightsource position information, the position information of the target object on the target picture, and the contour information. Through the method, the device and the electronic device, the light source position in the picture can be accurately and conveniently obtained, and other first-related factors are considered in the process of generating the shadow effect of the virtual object, so that theshadow effect of the virtual object is more realistic, and the immersion feeling of the user is improved.