16 results about "High speed video" patented technology

This disclosure presents embodiments of a model acceleration method, a video generation method, an apparatus, a device, a medium, and a product. One specific implementation of the method includes: determining a sample set and an original video generation model, wherein the samples in the sample set include reference images and audio; performing conditional distillation on the original video generation model based on the sample set to obtain a first video generation model, wherein the number of forward propagation steps of the first video generation model is less than the number of forward propagation steps of the original video generation model; and performing distribution matching distillation on the first video generation model based on the sample set to obtain a high-speed video generation model, wherein the number of inference steps of the high-speed video generation model is less than the number of inference steps of the original video generation model. This implementation relates to audio-driven video generation, reducing the computational complexity and inference latency of the model, and facilitating deployment and use in scenarios with limited computing resources.

Systems and methods transfer video data to an image processing system from a video source. Pixel data is received in a local buffer of a network interface controller and provided in a video transport packet. The video transport packet includes the pixel data, a media access control header and a video header. The video transport packet is received by another network interface controller that provides the pixel data directly into a video frame buffer of the image processing system.

The invention relates to the field of video collection, in particular to an anti-frame-loss high-speed video collection device and method, and the device comprises a video collection module which is used for collecting video data, adding a frame identifier for each video frame in the video data to form frame data, and transmitting the frame data to a video storage module; receiving a lost frame identifier sent by the video storage module, determining corresponding frame data according to the lost frame identifier to obtain lost frame data, and resending the lost frame data; the video storage module is used for receiving the frame data and the lost frame data, calculating a frame identifier of a lost video frame according to the frame identifier to obtain a lost frame identifier, and sending the lost frame identifier to the video acquisition module; and arranging the received frame data and the lost frame data according to the frame identifier to obtain video data. According to the invention, the problem of bandwidth preemption of the main data stream by real-time retransmission is avoided; therefore, the integrity of the video frame is ensured on the premise that the finally generated video stream does not interfere with real-time transmission at all.

The invention relates to the technical field of projection display, and discloses a vehicle-mounted ground projection method and system, a server, a terminal and projection equipment, a user terminal is adopted to send a video which a user wants to project, the server is adopted to encode the video, to-be-projected video data expressed in a bit plane format is obtained, and the to-be-projected video data is sent to the projection equipment. An executable program is determined according to to-be-projected video data represented in a bit plane format, a server replaces a graph generation step, the executable program is sent to a user terminal, the user terminal sends the executable program to vehicle-mounted ground projection equipment, the vehicle-mounted ground projection equipment stores the executable program in an internal memory, and the vehicle-mounted ground projection equipment stores the executable program in the internal memory. The vehicle-mounted ground projection equipment reads the executable program of the internal memory to the projection chip and performs projection display by using the projection light path, in the process, an external high-speed video interface is replaced, the video which is not stored in the internal memory of the vehicle-mounted ground projection equipment can be projected, and the user experience feeling is improved.

The invention designs a video data storage and processing integrated universal platform, which comprises a mainboard for receiving external video data, compressing the external video data, writing the external video data into an external storage device or an on-board storage space, and performing situation awareness or data fusion processing on the received video data according to requirements; the interface conversion board is used for carrying out interface expansion on the mainboard, connecting other interfaces except the interface of the mainboard, and converting the other interfaces into interfaces matched with the mainboard; and the external storage device is used for transferring the received video data to the external storage device when the storage space of the mainboard is not enough. The video data storage and processing integrated platform realizes the function of processing and storing various signals, especially high-speed video signals, on a single platform, has the characteristics of simplicity in operation, high transmission rate, wide application range and the like, greatly improves the applicability and reliability of the video data storage and processing integrated platform, and is suitable for popularization and application. The problems of single function, complex operation, small storage space, low transmission rate and the like of storage equipment used by an existing aerial bomb product are solved.

This invention discloses a phase correlation matching method for high-speed video measurement of small inter-frame offsets, comprising: acquiring a high frame rate image sequence; dynamically adjusting the image sequence through an adaptive window to obtain neighborhood sub-images; performing gradient real and imaginary dual-channel Gaussian filtering on the neighborhood sub-images to obtain a normalized cross-power spectrum matrix; performing plane fitting on the normalized cross-power spectrum matrix, and iteratively optimizing the sampling subset by dynamically adjusting the noise estimation index and inlier probability to obtain an initial offset value; based on the initial offset value, updating the normalized cross-power spectrum to calculate the residual offset value, iterating until the residual offset value is lower than a preset threshold, and summing all iterative offset values ​​to obtain the final sub-pixel offset value. This invention utilizes the robustness of frequency domain phase correlation matching to illumination changes and noise to achieve accurate identification and tracking of small inter-frame offset sequences of target points on the surface of high-speed moving objects under the influence of illumination, angle, etc., thereby ensuring the accuracy of dynamic monitoring.

The invention relates to the technical field of oily sewage treatment and separation, in particular to an oily sewage demulsification coalescence device and method based on ultrasonic field enhancement. The device comprises an ultrasonic generation system, an oil drop coalescence dynamic control system, a high-speed video acquisition system, a water bath constant temperature system and a liquid level control system, the oil drop coalescence dynamic control system is used for forming a controllable flow shear field and an oil drop system in a micro-flow test container by utilizing a micro-flow control technology; the ultrasonic generation system enables a controllable ultrasonic field to directly act on the flow system, and the field coupling effect under the real working condition is simulated; the high-speed video acquisition system is used for carrying out high-temporal-spatial-resolution imaging and recording on deformation, movement and coalescence processes of oil drops in the coupling field; and the water bath constant temperature system and the liquid level control system ensure the stability and repeatability of experimental conditions. All the systems work cooperatively, and in-situ, quantitative and visual observation of the whole ultrasonic demulsification process is achieved under the condition of simulating real flow.

The application relates to a long-distance continuous pose high-speed video measurement method, which comprises the following steps: step 1, using a multi-biocular continuous observation joint bundle method adjustment to solve the observation value; step 2, using a self-adaptive variable weight strategy of depth distance and adjustment residual error constraint to process the solving process of step 1), so as to obtain a continuous pose parameter measurement result; and step 3, using a position and posture parameter dynamic inversion based on plane fitting to optimize the continuous pose parameter measurement result obtained in step 2). Compared with the prior art, the application has the advantages that long-distance moving target dynamic position and posture parameters can be measured continuously with high precision, the measurement positioning precision is better than 3mm, and the precision is improved by more than 30% compared with the precision of an international famous visual measurement software.

This invention relates to the field of KVM switching control and intelligent video transmission, specifically to a seamless switching control system and method for KVM switches based on reinforcement learning. The system includes: collecting physical layer state data of the current data link and the target data link, including cable impedance attenuation values, bit error rate before adaptive equalizer compensation, historical timing data readout time, estimated environmental signal-to-noise ratio, and terminal configuration mode; inputting the physical layer state data into a reinforcement learning decision engine based on a deep Q-network, outputting link compensation parameters and pre-loaded timing parameters; performing a virtual handshake operation based on the pre-loaded timing parameters to generate a source-end link hold state; completing the reconstruction of underlying physical layer parameters and cross-matrix switching in the vertical blank area of ​​the video stream, and updating the model reward value based on the retrained state of the switched link; this invention achieves adaptive seamless switching control for high-speed video links such as display ports.

The invention discloses a lossless high-speed camera link video signal rotation transmission system which comprises a fixed disc, a transmitter, a receiver, a transmission shaft, a bearing end cover, an optical fiber slip ring, an optical fiber chuck, a rotor shaft, a stator shaft, a rotor bearing, a stator bearing and a cover plate. Input and output signals of the video signal photoelectric rotation transmission system need to be scrambled and descrambled through the transmitter and the receiver, and the reliability and stability of the input and output signals are ensured by encoding and decoding the input and output signals. The conductive slip ring has the advantages of low time delay, long service life, high signal transmission reliability, strong stability and capability of ensuring the reliability and the stability of signal transmission, and solves the problems of short working period, difficulty in maintenance, unstable / distorted signal transmission and the like when a common conductive slip ring is used for transmitting a high-speed video signal.

The invention discloses a vehicle converse driving detection method for a highway, and the method comprises the steps: reading single-frame initial data of a high-speed video stream, inputting the single-frame initial data into a lane line detection model, obtaining a lane line, and generating initial lane information; adaptively generating a first region and a second region; combining the two target detection areas, inputting the two target detection areas into a target detection model to obtain vehicle detection results, fusing the vehicle detection results, and inputting a fused result into a multi-target tracking algorithm to obtain vehicle track information; judging the driving direction of the lane; and based on the track information of the current vehicle and the driving direction of the lane where the current vehicle is located, identifying whether retrogradation occurs. By analyzing the relative position relation between the vehicle movement track and the identified lane in real time, the problem of high calculation overhead caused by analyzing the lane area frame by frame in a traditional high-speed scene vehicle converse driving detection method can be solved.

The utility model provides a beverage production line bottle falling detection device based on visual detection, which is arranged at a bottle falling position of a beverage production line and comprises a movable support, a high-speed camera, a high-speed video camera and a computer, the high-speed camera, the high-speed video camera and the computer are arranged on the movable bracket; the high-speed camera and the high-speed video camera are connected with the computer; when a bottle falling phenomenon occurs in the beverage production line, a maintainer operates the computer to play back pictures recorded by the high-speed camera so as to determine the reason of bottle falling. The high-speed camera is used for recording the whole process of bottle transferring pictures, and the dynamic process of the bottle falling moment can be collected. Compared with the prior art which can only judge the bottle toppling result, the device can provide frame-by-frame backtracking high-definition images for maintenance personnel to help the maintenance personnel to visually analyze bottle toppling inducements such as abnormal speed of the conveying belt, surface structure defects and guardrail scratching, and the accuracy and efficiency of fault diagnosis are greatly improved.

The utility model discloses time interval measuring equipment based on a high-speed video collector, and belongs to the technical field of time interval measurement and metering traceability. In order to solve the technical problems of high measurement uncertainty and incapability of accurately tracing due to difficulty in synchronous triggering of visual triggering type timing equipment, the utility model comprises at least two independently adjustable lenses and light paths, the light guide system is used for receiving and deflecting light of each light path; and a high-speed photosensitive device. The key point is that the light guide system converges light from different lenses and images the light on different areas of the single high-speed photosensitive device respectively. Visual information of a time standard source and a detected object can be synchronously recorded at a high speed through the photosensitive device, and an accurate time interval is calculated by analyzing a video key frame and combining with a known frame rate. According to the utility model, the synchronization error is effectively eliminated, the uncertainty is reduced to millisecond level, the high-precision measurement and traceability of visual trigger equipment are realized, the structure is compact, and the adjustment is convenient.

The invention discloses a method for analyzing and processing a high-speed video sequence in a liquid drop collision process, which comprises the following steps of: shooting the liquid drop collision process through an experiment, creating a layered output directory according to a video image, automatically modeling a background and preprocessing a video, detecting a key event and correcting the image, and analyzing and calculating liquid drop collision kinetic parameters. And verifying a result and outputting data. The method has the advantages that real end-to-end full automation is achieved, the method has the capacity of processing mass video data, the precision and robustness are improved in a leap-over mode, the most stable and most recognizable features are found for calculation by means of the dynamic characteristics of liquid drop impact, subjective errors caused by manual point selection are fundamentally eliminated, and the accuracy and the accuracy of point selection are improved. According to the method, the correction precision is far higher than that of a traditional method, remarkable synergistic effects and technical effects are generated, objectivity and repeatability of data are ensured, reliability and authority of scientific research data are greatly enhanced, and a solid technical guarantee is provided for scientific research.

The application provides a DVI video signal transmission method and system with adaptive equalization parameter correction. A feature code or a control word is added to a DVI signal transmitting end, and the form of the feature code or the control word is analyzed and recognized at a receiving end to diagnose whether the received video quality meets the display requirement. When the received video quality does not meet the display requirement, the related coefficient of the receiving end is adjusted by adaptive optimization of equalization parameters, so that high-quality display of the received video is realized. The application realizes self-defined video quality diagnosis technology to determine the video transmission quality, and automatically adjusts the equalization parameters of the receiving end by the method of optimization of equalization parameters, so that the problem that the equalization parameters need to be manually adjusted due to the inconsistency of the characteristics of different transmission cables in the process of high-speed DVI video transmission is solved, and the transmission compatibility of high-speed DVI video signals is improved.