48 results about "Motion magnification" patented technology

The invention discloses a flexible electrothermal drive micro-gripper and a manufacturing process method, belongs to a micro-actuator in the technical field of micro-electro-mechanical systems and in particular relates to a flexible electrothermal drive micro-gripper and a manufacturing process method thereof. The micro-gripper is of an integral structure with two parts (a driving part and a flexible mechanism of a gripper body) and comprises an SU-8 adhesive structure layer, an upper electrode metal layer and a lower electrode metal layer; and furthermore, the left structure and the right structure are symmetrical completely. An inverted V-shaped beam array is utilized as the driving part, and the part of the flexible mechanism of the gripper body is a 10-rod flexible mechanism. The manufacturing process method comprises the steps of oxidation, adhesive coating, exposure, development, corrosion and sputtering of metal. The temperature at the gripping end of the micro-gripper is lower, the three-layer symmetrical structure is adopted, and bending in normal direction can be avoided during the gripping process. The flexible mechanism in the micro-gripper has a novel structure, the motion magnification factor is large, and the range of sizes of gripped objects is wide; and furthermore, due to the integral structure, the assembly is unnecessary, the manufacturing is simple, and the micro-gripper and the manufacturing process method are conductive to realizing the practicality of a biological micro-operation technology.

The embodiment of the application discloses a bridge vibration detection method and related devices, which are applied to vibration detection equipment and include a camera device. The bridge vibration detection method comprises the following steps: collecting an original video of a bridge to be detected through the camera device, processing the original video according to a preset video processing strategy to obtain a target video with a motion amplification effect, and determining reference vibration data of the bridge to be detected according to the target video, wherein the motion magnification effect refers to that the area where the bridge to be detected moves is magnified in the target video; determining an operating mode of the bridge to be detected according to the reference vibration data; obtaining a pre-stored mode of the bridge to be detected; comparing the pre-stored mode and the operating mode to determine if the vibration state of the bridge to be detected is abnormal,wherein the pre-stored mode is the mode corresponding to a designed characteristic frequency of the bridge to be detected; and if so, outputting preset information. The bridge vibration detection method and related devices are beneficial to non-contact detection of a bridge vibration state, and improve the effectiveness, convenience, and versatility of vibration detection.

The invention provides a blood pressure detection method for extracting pulse waves based on videos and a terminal. The method comprises the steps that parameters of a nonlinear blood pressure calculation model are initialized; using a camera to shoot videos of neck and facial artery pulsation regions of the subject; amplifying a weak neck artery pulse signal and a weak facial artery pulse signal in the two regions by using a motion amplification technology; respectively selecting neck and facial artery pulsation point regions with obvious pulsation, and calculating to obtain pulse wave conduction time between the two point regions; a nonlinear blood pressure calculation model is adopted, and the average blood pressure value of the subject in the shooting time is obtained through calculation. According to the method, the neck and facial artery regions are collected through the camera, the pulse wave conduction time of the neck and facial artery pulsation points is obtained through signal amplification extraction and calculation, finally, the blood pressure value is obtained through calculation in combination with a nonlinear blood pressure calculation model, and non-invasive, non-contact and non-inductive blood pressure measurement is achieved while time delay errors are avoided.

The invention discloses a bridge monitoring method, system and device based on a tiny vibration amplification technology, and the method comprises the steps: decomposing a vibration video image of a to-be-monitored bridge into baseband images with different frequencies, and separating out the vibration images of the to-be-monitored bridge from the baseband images with different frequencies; the method comprises the following steps: respectively carrying out motion amplification on first N orders of vibration signal images in vibration images of different frequency base bands of a to-be-monitored bridge, so that tiny vibration which cannot be directly monitored originally in a vibration video image can be monitored; and superposing the front N-order vibration signal images after motion amplification and the vibration signal images which are not amplified to obtain reconstructed vibration video images of the bridge to be monitored, in which the front N-order vibration signal images are amplified, and realizing monitoring of micro vibration displacement of the bridge to be monitored according to the reconstructed vibration video images. Therefore, monitoring of bridges with very small vibration amplitudes, such as small and medium span bridges, is realized, additional arrangement of sensors is not needed, and the cost is reduced.

The embodiment of the application discloses a method for detecting abnormality of a gear set and a related product, and is applied to vibration detection equipment. The method comprises the followingsteps: an original video of a target gear set arranged in a transmission of a motor vehicle in an operating state is obtained; the original video is processed according to a preset video processing strategy to obtain a target video with a motion magnification effect; reference vibration data of the target gear set is generated according to the target video; and abnormal causes of the target gear set are determined according to the target video and/or the reference vibration data. The embodiment of the application reduces the detection difficulty when the gear set is abnormal, and improves theefficiency for detecting the abnormal causes of the gear set.

In one process, a mechanical stress (quasi-static or dynamic) or thermal stress is introduced into a local area of a structure where damage is suspected. A standard (or high speed for some dynamic stress approaches) video recording is made of the area under test. The video is processed using a motion amplification algorithm which magnifies extremely small motions. The motion amplification can reveal strains (for static or thermal loading) or wave patterns (for dynamic loading) generated on the surface of the part. If there is damage or structural change, the resulting surface strains will indicate it, or the surface waves in that area will change in phase, amplitude, or both, and will be indicated in the video images. In another process, motion magnification is used to highlight local failure events during structural testing.

The invention provides a non-contact body temperature and respiration rate combined detection method. The method comprises the steps of acquiring video image data under the target background of a human body by a thermal imager or a night vision device, wherein the video image data comprise temperature field information in a scene, and the temperature of the human body can be directly read in the thermal imager or the night vision device; amplifying the acquired video image data by using a micro motion amplification technology; and carrying out optical flow calculation on the amplified video image data to finally obtain an accurate value of respiration rate. According to the method, infrared video images are captured by the thermal imager or the night vision device, the body temperature andthe respiration rate of the crowd are jointly detected, and even under the condition that the body temperature is normal due to the fact that a specific target crowd takes antipyretics, the specifictarget crowd can be identified through the characteristic change of the respiration rate of the specific target crowd and an alarm is given.

The invention discloses a human face living body detection system applying a double-branch three-dimensional convolution model, a terminal and a storage medium, and belongs to the field of human faceliving body detection. The system comprises a face video acquisition module, a face video preprocessing module, a living body labeling module, a living body motion amplification module, a double-branch three-dimensional convolution living body recognition training module and a living body judgment module. The double-branch three-dimensional convolution model training module is configured with a double-branch three-dimensional convolution model and comprises a static texture information sub-module, a dynamic motion clue sub-module, a fusion sub-module and a classification sub-module; the outputs of the static texture information sub-module and the dynamic motion clue sub-module are pooled, summarized and fused through the fusion sub-module, and then a detection result is output through theclassification sub-module; the invention is mimicry and has biological significance, and has very strong robustness and generalization; a great living body detection guarantee can be provided for a face recognition system, the safety of the system is improved, and information and properties are protected from being damaged.

The invention belongs to the field of micro driving application and relates to an object lens driving table and specifically to an object lens driving table using a hinge mechanism to magnify output displacement. The object lens driving table is used for fixing an object lens ring and comprises a hollow housing and an actuating member disposed on the housing. The hollow housing comprises a housing portion and a chamber passing through the housing portion. The hinge mechanism is arranged on the housing portion and comprises a moving portion and a moving platform connected with the moving portion. The actuating member is abutted against the moving portion. The object lens ring is arranged in the chamber and is fixed to the moving platform. According to the object lens driving table, the hollow housing is provided; the hinge mechanism is disposed on the housing; the object lens ring is disposed on the moving platform of the hinge mechanism; and the actuating member is abutted against the moving portion of the hinge mechanism. Thus, the hinge mechanism magnifies the motion transmitted by the micro displacement of the actuating member in a gapless and uncoupled manner so as to effectively increase a motion magnification factor and guarantee size compactness.

The invention discloses a petrochemical gas leakage detection method, system and device and a storage medium, and relates to the technical field of environment monitoring, and the method comprises the steps: obtaining an original infrared video of petrochemical gas leakage, processing the original infrared video through a preset motion amplification method, and obtaining a target video; extracting a leaked image frame in the target video, and establishing a data set based on the leaked image frame; constructing an initial petrochemical gas leakage detection model, and training the initial petrochemical gas leakage detection model by using the data set to obtain a target petrochemical gas leakage detection model; and obtaining a to-be-detected video subjected to motion amplification processing, inputting the to-be-detected video into the target petrochemical gas leakage detection model to obtain a detection result, and judging whether petrochemical gas leakage exists in the to-be-detected video according to the detection result. The technical effect of improving the petrochemical gas leakage detection accuracy is achieved.

The invention discloses a method for accurately extracting a pulse wave based on a human body superficial artery video, which is used for solving the problems of movement and illumination interference generally existing in a non-contact pulse wave measurement method and carrying out multiple filtering on various noises to realize accurate extraction of the pulse wave. The method comprises the following steps: S1, acquiring a human body superficial artery video; s2, preprocessing the video through adaptive gamma transformation; s3, motion amplification is carried out on the video through an amplification algorithm combining Euler amplification and FRR filtering; s4, an adaptive threshold inter-frame difference method is adopted, and a multi-feature region is selected as a region of interest; s5, extracting an original pulse wave signal; s6, carrying out the wavelet transformation of the II-type Chebyshev method and the adaptive threshold method, and filtering the noise in the signal through an adaptive notch filter; and S7, enhancing the pulse wave signal by a modulation domain spectral subtraction method. According to the method, the influence of motion artifacts and ambient brightness differences can be effectively inhibited, meanwhile, the region of interest is accurately selected to filter noise, and the accuracy of pulse wave extraction is further improved.

The invention discloses a concrete arch dam modal shape recognition method based on unmanned aerial vehicle video shooting. The method comprises the steps of arranging mark points at the dam crest; obtaining a dam crest image through a pan-tilt camera on the unmanned aerial vehicle, and obtaining a dam crest vibration video; according to the first four-order natural frequency of the arch dam, obtaining a video after the first four-order motion of the original dam vibration video is amplified; and selecting a frame with the maximum dam crest deformation in the original four-order motion amplification video of the dam, identifying the centroid coordinates of each mark point in the image, obtaining the centroid coordinate change of each mark point before and after vibration, and normalizing the maximum value of the relative coordinate change of each mark point to obtain the four-order modal shape of the dam crest of the arch dam. Through a non-contact method, approximate full-field measurement of the arch dam crest vibration mode can be realized, the spatial resolution is high, and support can be provided for subsequent dam model updating and damage identification.

The invention provides a non-contact torsional vibration measurement method based on machine vision. The method comprises the steps of performing high-speed video acquisition on a torsional vibrationobject by using a computer vision technology, realizing separation, extraction and amplification of torsional vibration signals by using a 'micro motion amplification 'principle, and calculating the instantaneous frequency of the torsional vibration signals by using Fourier transform, wavelet transform and other technologies. According to the method provided by the invention, the torsional vibration time domain signals can be accurately extracted, and the torsional vibration frequency and period can be obtained. And the method is still effective to a tiny torsional vibration system. The methodhas the characteristics of being non-contact, high in test efficiency and only short in video acquisition time. Only a camera and lighting equipment are needed, a sensor and a data acquisition systemare not needed, mark points do not need to be pasted on the surface of the torsional vibration object, and accuracy and reliability are good. According to the method provided by the invention, the state of the shaft system of a rotary machine can be monitored. For a turntable system or a slender structure, the rotational inertia of the system, the shear modulus of the material and the like can becalculated.

The invention provides a response measurement method for an aircraft model in a wind tunnel. The method comprises the steps: pasting marking points on an aircraft wind tunnel model, employing a high-speed camera for continuously recording a video, directly observing buffeting and flutter frequencies under the conditions of micro vibration, high-frequency vibration and large amplitude through a micro motion amplification algorithm, further carrying out the optical flow calculation on the amplified video, and measuring an attack angle; enabling the Mach number to remain as a constant, rotating the aircraft model to enable the attack angle to change continuously in a large range, if the pulsation frequency and the amplitude in attack angle measurement data are suddenly changed, determining that the suddenly changed angle is the initial attack angle of buffeting, and the corresponding lift coefficient is the initial lift coefficient of buffeting; after buffeting occurs, performing signal enhancement and filtering on the mark points on the aircraft model by using a micro motion amplification algorithm, so non-contact measurement and extraction of vibration signals can be realized. The method has the advantages of being high in reliability and precision, free of contact, easy to implement and the like.

The invention discloses a non-contact mental stress evaluation system, which comprises a video acquisition module, a data processing module and an analysis and evaluation module; the video acquisition module is used for acquiring an upper body video of the to-be-tested person; the data processing module is used for carrying out feature point detection on the divided face region and identifying the micro-expression of the person to be detected, performing color amplification on the divided face area by utilizing Euler color amplification, and extracting a pulse wave signal to obtain heart rate and heart rate variability physiological parameters, acquiring the displacement signal and the breathing rate of the chest wall through Euler motion amplification, performing data dimension reduction on the extracted micro-expression, heart rate and heart rate variability physiological parameters, the displacement signal of the chest wall and the respiratory rate by using a data dimension reduction technology; an analysis and evaluation module which is used for analyzing and evaluating the data output by the data processing module by using a non-contact mental stress evaluation model, or training the non-contact mental stress evaluation model. The method has the advantages of few sensors, simplicity in operation and low cost.

The invention discloses a micro-expression recognition method based on adaptive motion amplification and a convolutional neural network. The method comprises the following steps: 1, converting a certain sample of a micro-expression video into an image sequence, and carrying out the face cutting and alignment; 2, reading an initial frame of the image sequence, and calculating by using a vertex frame positioning algorithm to obtain a vertex frame picture; 3, determining a proper magnification factor by adopting a self-adaptive motion magnification method, and performing motion magnification on the vertex frame according to the determined magnification factor so as to enhance the characteristics of the micro-expression; 4, obtaining optical flow characteristics of the micro-expression video according to the start frame and the amplified vertex frame, and obtaining a horizontal optical flow, a vertical optical flow and optical strain; 5, establishing a convolutional neural network model for micro expression recognition, and performing transfer learning from macro expression to micro expression by using the model; and 6, inputting the optical flow features into the model after transfer learning, outputting time-space features, and training the model to realize micro-expression recognition.

Example embodiments allow for fast, efficient motion-magnification of video streams by decomposing image frames of the video stream into local phase information at multiple spatial scales and/or orientations. The phase information for each image frame is then scaled to magnify local motion and the scaled phase information is transformed back into image frames to generate a motion-magnified video stream. Scaling of the phase information can include temporal filtering of the phase information across image frames, for example, to magnify motion at a particular frequency, ln some embodiments, temporal filtering of phase information at a frequency of breathing, cardiovascular pulse, or some other process of interest allows for motion-magnification of motions within the video stream corresponding to the breathing or the other particular process of interest. The phase information can also be used to determine time-varying motion signals corresponding to motions of interest within the video stream.

The invention provides a Lagrange video motion amplification method based on an image deformation technology. The method comprises the steps: estimating the motion information of feature points of a region of interest in a video through an optical flow method, amplifying the motion displacement, and then taking the coordinates of the feature points before and after amplification as traction points, and achieving the motion amplification of an image through an image deformation technology for maintaining image content. Compared with the current motion amplification technology, the method has less blurring and ringing artifacts, and each frame of image can be processed only by a small amount of manual intervention. An input picture sequence is from a static camera; an amplification area canbe manually selected and amplification times can be set, the whole system can automatically detect and track feature points, a target area is deformed through an image deformation technology after motion is estimated again, and finally, a foreground and a background are reconstructed to obtain an amplified video sequence, so physical motion of a target in a video is amplified.

The present disclosure provides a washing machine vibration fault detection method and device. The method comprises the following steps of determining a standard vibration mode according to a vibration detection type option selected by a user; acquiring a target video corresponding to a washing machine; amplifying the target video by adopting a Lagrange motion amplification method to obtain an amplified output video; acquiring image frames corresponding to the amplified output video, and calculating cross power spectrum between image frames by adopting a phase correlation algorithm for the image frames; performing inverse Fourier transform on the cross power spectrum to obtain vibration information of pixels in the target video; and matching the vibration information with the vibration information corresponding to the standard vibration mode, and determining the vibration fault of a vibrating object according to the matching result. By adopting the embodiment of the application, the Lagrange motion amplification method is adopted to amplify the micro motion in the video, and then the vibration information in the video is extracted, so that the accuracy of information extraction isimproved, and the reliability of vibration analysis is further improved.

The invention relates to a video motion amplification method based on an improved self-encoding network, and uses the improved self-encoding network to amplify subtle changes in a video. The method comprises the following steps: decomposing video data, and adopting two continuous frames of decomposed images as the input of an encoder; using an encoder to extract shape features of two continuous frames of images as input of an amplifier; amplifying the pixel displacement difference value of the shape features of the two frames of images by using an amplifier to obtain amplified shape features; and performing up-sampling on the texture features of the previous frame image by using a decoder, and combining the texture features with the amplified shape features to obtain an amplified frame and output the amplified frame. According to the method provided by the invention, perfect fusion of shape and texture features in the video motion amplification image is realized, the brightness, color and texture loss of video motion amplification is reduced, and shallow feature information is reserved.

The invention provides a high-frequency vibration mode identification method and system based on motion amplification and digital image correlation, and solves the problems that an existing vibration mode test mode is inconvenient to install and low in positioning precision; the test process is slow and low in efficiency; and the high-order vibration mode test accuracy is low. The method comprises the following steps: 1) fixing a measured structural member and spraying speckles on the surface of the measured structural member; 2) arranging a laser Doppler vibration meter on the outer side of the measured structural member; 3) applying an excitation signal to the measured structural member, and triggering the laser Doppler vibration meter to acquire a response speed signal of the measured structural member; 4) processing the excitation signal and the response speed signal by adopting an estimation method to obtain a frequency response function, and finding out outburst frequency points from the frequency response function and representing the outburst frequency points as f1, f2,..., fn; (5) sine excitation signals with the fixed frequencies of f1, f2... fn are applied to the measured structural component, and meanwhile, a shot video under each sine excitation is shot; (6) motion amplification is carried out on each shot segment of the shot video; and 7) performing DIC analysis on the photographic video after motion amplification to obtain a high-frequency vibration mode.