21results about How to "Low hardware requirements" patented technology

The application discloses a visual perception method for assisted driving, comprising the following steps: S100, collecting assisted driving image data, and constructing input X after data set preprocessing; S200, constructing a multi-task neural network model F and substituting the input X into the model to obtain a plurality of branch task-specific outputs f i ; S300, calculating the final loss function L all of the multi-task neural network model F; S400, optimizing the minimum as the target of the loss function L all , and finally obtaining the optimal multi-task neural network model F * . A system using the above visual perception method for assisted driving is also disclosed, specifically comprising a data input module, a shared layer feature extraction module, a task layer structure search module and a model output module. The application has the beneficial effect that a multi-task model is constructed to effectively reduce the network parameter quantity and the calculation quantity, simultaneously reduce the online reasoning time, and is very beneficial to practical application.

The application provides a display device, which comprises a liquid crystal prism panel, the liquid crystal prism panel comprises a plurality of electrode groups, the electrode groups comprise a plurality of electrodes arranged in sequence and at intervals; the absolute values of voltages on at least two electrodes in the electrode groups comprise a first voltage segment, and the proportion of the time length of the absolute values of voltages on the at least two electrodes in the first voltage segment in the time length of the first time segment is different. The display device provided by the embodiment of the application reduces the complexity of the driving circuit of the liquid crystal prism panel and reduces the hardware requirements for the driving circuit of the liquid crystal prism panel.

This invention relates to the technical field of seismic data interpretation, and discloses a method and computer equipment for identifying the locations of beaded reflections in carbonate rocks. The identification method includes: S1, obtaining a two-dimensional seismic migration profile; S2, extracting typical beaded reflections, determining their dimensions, statistically analyzing their structural morphology, and constructing adaptive hit-and-miss structural element pairs; S3, using a hit-and-miss transformation to identify locations that match the adaptive hit-and-miss structural element pairs. This invention's identification method extracts multiple typical beaded reflections, constructs adaptive hit-and-miss structural element pairs, and only identifies locations that match these pairs. Based on a hit-and-miss transformation algorithm, this algorithm has low hardware requirements, high computational efficiency, and high accuracy.

The embodiment of the application relates to the radar technical field, and discloses a signal processing method, an electronic device and a storage medium. The signal processing method comprises the following steps: transmitting a detection signal in a Doppler multiplexing (DDM) mode through a plurality of transmission channels; processing a return signal to obtain distance-Doppler two-dimensional data; calculating a likelihood probability of a Doppler spectrum under different aliasing combination conditions of transmission channel orders of at least two targets according to the distance-Doppler two-dimensional data; and determining Doppler spectra of the targets according to an aliasing combination corresponding to a maximum value of the likelihood probability. At least, it is beneficial to correctly restore signals before Doppler spectrum aliasing in the presence of at least two targets.

The application discloses a three-level coordinated production line body intelligent control method and system, and belongs to the technical field of intelligent manufacturing. The method comprises the following steps: establishing a global task planning layer, performing semantic analysis and logical disintegration at a first preset frequency, and generating an atomic action sequence; establishing a local space perception layer, receiving the atomic action sequence at a second preset frequency, constructing a dynamic local map by using sensor data, determining a workpiece pose by point cloud registration, and mapping the atomic action sequence into a continuous motion trajectory flow; establishing a real-time action execution layer, receiving an instantaneous target pose at a third preset frequency, calculating a theoretical driving torque by combining a Lagrange dynamics equation, switching to an impedance control mode at a contact moment, and driving an execution mechanism to complete work; and establishing an asynchronous coordination layer, realizing three-layer asynchronous communication through an industrial network, and controlling the three-layer frequencies to be distributed in a ladder type in an increasing manner. The application has the effects of improving the flexibility, precision and safety of a production line.

The application discloses an online action detection method based on a light-weight Transformer, and comprises the following steps: setting a detection model, wherein the detection model is based on a Transformer model; acquiring features of a video stream with a length of T; inputting the features of the video stream into the detection model; and outputting an action category of the last frame of the video stream through the detection model as a detection result. The online action detection method based on the light-weight Transformer disclosed by the application reduces the calculation amount while improving the accuracy.

Embodiments of the present specification disclose an image defogging method and device, electronic equipment and storage medium. The method comprises acquiring an original format image, and performing exposure gain compensation on the original format image to obtain a color reconstruction image. Then, the R channel component of the original image is extracted to obtain a first image corresponding to the R channel component. Further, the color reconstruction image is decomposed based on a wavelet transform algorithm to obtain a contour layer image and a detail layer image. Finally, the first image and the contour layer image are fused to obtain a fused image, and the fused image and the detail layer image are processed by inverse wavelet transform to obtain a defogging image. Through the above image processing steps, the object signal intensity in the finally obtained defogging image is high, so that even in a heavy fog shooting environment, the defogging method can achieve an ideal defogging effect, and the method has low complexity, good real-time performance, low requirements for device hardware and computing power, and meets the requirement of high applicability.

The invention discloses a fast regional self-adaptive ACM method based on meta-learning, and belongs to the technical field of low-orbit satellite communication. Aiming at the technical problems of dependence on a large amount of local data, long debugging period, unstable performance and the like during cross-region deployment of the existing adaptive coding modulation technology, the method comprises the following steps of: constructing a meta-training task set covering various global climate characteristics, and training to obtain a meta-initial model with strong generalization ability; when a ground station is deployed in a new region, an optimized localized ACM strategy can be quickly adapted through several steps of gradient updating by using a very small amount of initial communication data collected by the station. Simulation results show that the method only needs 30 samples and 2-minute fine tuning to achieve the approximate optimal performance, the spectrum efficiency is improved by 35.8% compared with a traditional fixed threshold value method, the sample demand is reduced by 99% and the deployment time is shortened by more than 98% compared with a supervised learning method, the regional self-adaption problem in global rapid deployment of the satellite communication system is effectively solved, and the method has good application prospects. And the operation cost is obviously reduced.

The application discloses a ranging method, comprising: obtaining a mapping relationship between serial numbers of actual data boxes and serial numbers of preset data boxes, wherein each actual data box is mapped to at least one preset data box; obtaining ranging data, assigning storage space to the actual data boxes and storing the ranging data; determining a target data box and a serial number of the target data box according to data distribution of the ranging data in the actual data boxes; determining a target value of a preset data box corresponding to the target data box according to the mapping relationship between the serial numbers of the actual data boxes and the serial numbers of the preset data boxes and according to a preset algorithm; and determining a distance of a measured target according to the target value of the preset data box and a width of the preset data box. The application also provides a ranging system, which can reduce storage space.

This invention discloses a step-by-step synthesis method for nodal instrument shot gather data, relating to the field of seismic exploration technology. The invention includes the following steps: collecting source point shot-receiver relationship files for synthesizing single-shot data to form a general source point shot-receiver relationship file; determining the start and end range of the general source point shot-receiver relationship file's file numbers based on the single-shot data file numbers corresponding to each source point in the general source point shot-receiver relationship file; and identifying the minimum and maximum station numbers among all receiver alignments based on the general source point shot-receiver relationship file to determine the alignment line station number range. This invention employs a step-by-step synthesis method. After the receiver alignment is retrieved, all data collected by that alignment can be segmented and synthesized for output, advancing the shot gather data synthesis time. This allows for earlier understanding of the nodal alignment data situation, and multiple machines from multiple nodal data synthesis working groups can perform the data synthesis simultaneously, solving the problem of existing nodal instrument shot gather data being difficult to synthesize step-by-step, resulting in delayed data presentation.

The application discloses a kind of detection method, device, readable medium and electronic equipment of audio playing equipment, the method comprises: obtaining the audio data to be measured played by the audio playing equipment to be measured;According to the preset segmentation duration, the audio data to be measured is divided, and a plurality of audio segments to be measured are obtained;Extract the audio feature to be measured of audio segment to be measured, and the audio feature to be measured is clustered and predicted, and the prediction result of audio segment to be measured is obtained;According to prediction result, matching is carried out in preset classification library, to determine the predicted category of audio segment to be measured;According to the predicted category of a plurality of audio segments to be measured, the detection result of audio playing equipment to be measured is determined, and detection result includes normal equipment and abnormal equipment.The application embodiment can be applied to cloud technology, artificial intelligence, intelligent transportation, auxiliary driving and various scenes.The technical scheme provided by the application embodiment refines detection data, improves detection precision and accuracy, and has fast calculation speed, effectively improves detection efficiency.

The application relates to a manta ray robot fish roll control method based on Q learning and pectoral fin amplitude bias, compared with a traditional control method, the application does not need to establish a control object model, collects experimental data, trains a Q table offline, then transplants the table into a prototype controller, obtains a control variable through table lookup, and controls the prototype to swim at a fixed depth. Compared with other reinforcement learning control methods, the application has low requirements on hardware resources, small space demand, low power consumption and convenient realization. The application has the beneficial effects that the application has low requirements on the hardware of a controller, and an ordinary single-chip microcomputer can be used for realization, thereby saving cost, space and energy. The application does not need to establish a mathematical model of a control object, can save a large amount of time cost, and is difficult to establish a bionic model. The application can realize roll control of the robot fish without expert experience for establishing a rule base, and the Q table can learn "expert experience" during training.

The application discloses an image color style adjustment method and device based on an ISP. A source image is acquired from a first imaging system, and first probability density cumulative distribution data of the source image is determined; second probability density cumulative distribution data of a target image is acquired in advance, or the target image is acquired from a second imaging system, and the second probability density cumulative distribution data of the target image is determined; a gray scale mapping table of each component channel of the source image and the target image is obtained according to the first and second probability density cumulative distribution data; a third gray scale value of each pixel after adjustment is determined according to a first gray scale value of each pixel of each component channel of the source image and a corresponding gray scale mapping table, and an output image after color style adjustment is obtained. The color difference of different devices can be eliminated, an image with consistent color style is obtained, and the method has strong universality.

According to the LOD construction method and device for the GIS large-range live-action model, the nodes are generated through adaptive iterative segmentation in the HLOD tree construction process, the triangular patch data amount of each node is kept within the threshold value range, and the nodes with the too large triangular patch number cannot be generated; meanwhile, the textures of all the nodes are limited to be kept in a similar range, too large textures cannot be generated, and the optimal configuration effect can be achieved in the rendering scheduling process of the result; meanwhile, after the triangular patches in all the nodes are simplified, the texture features in the texture pyramid are merged, and the method can effectively simplify the grids, maintain the features of the grids, reduce the deformation of the textures to the maximum extent, and ensure the visual effect of the grids. In addition, main memory occupation data is stored in a temporary file by using a memory mapping technology, so that the scheme can process large data volume data exceeding computer memory limitation.

This invention discloses a visual detection method for evaluating the standardization of drainage in filtration operations. By employing a state machine dynamic triggering mechanism, the optimal observation window is defined as the stable existence of the funnel and the absence of glass rod obstruction. The method accurately extracts and caches the single-channel binary mask of the three filter paper layers to avoid real-time segmentation noise. Then, based on the mask, the orientation of the three filter paper layers and the contact position between the glass rod and the filter paper layers are determined in real time. Relying on two-dimensional conditions, the method identifies when the glass rod is ready. Real-time interactive analysis and timed segmentation are implemented to achieve extremely high real-time response performance and reduce hardware requirements. Finally, the method calls the opposite-side drainage rules to adjudicate compliance, achieving full-process automation. This method ensures segmentation accuracy and judgment accuracy while reducing real-time computing power consumption, efficiently solving the core problems of existing technologies and providing reliable technical support for evaluating the standardization of experimental operations.

The application relates to a manta ray robot fish heading control method based on Q learning and pectoral fin phase difference. Compared with a traditional control method, the application does not need to establish a control object model, collects experimental data, trains a Q table offline, then transplants the table into a prototype controller, obtains a control variable through table lookup, and controls the prototype to swim at a fixed depth. Compared with other reinforcement learning control methods, the application has low requirements on hardware resources, small space demand, low power consumption and convenient realization. The application has the beneficial effects that the application has low requirements on the hardware of a controller, and an ordinary single-chip microcomputer can be used to realize the application, thereby saving cost, space and energy, saving a large amount of time cost, being simple to realize, having strong generalization ability, and being applicable to not only a bionic underwater robot but also a traditional underwater AUV.

The invention provides a laser mass ball mounting method and device, and the method comprises the steps: providing through hole glass which is provided with a plurality of through holes penetrating through the upper surface and the lower surface, the upper part of each through hole is a big-end-up conical hole, each through hole bears a solder ball, and the plurality of through holes are arranged in a patterned manner; laser irradiates from the upper side of the through hole glass to the upper surface of the through hole glass, the solder balls in the through holes are melted, and the melted solder balls are transferred to a substrate below the through hole glass under the inert gas pressure; and a certain distance is formed between the substrate and the through hole glass. By the adoption of the method, the tin balls after laser melting are subjected to transfer type ball planting under the air pressure, the tin balls do not make contact with the substrate in the whole manufacturing process, the ball planting and welding yield is greatly increased, and the substrate material is not limited by the temperature.

The invention relates to a water quality monitoring data analysis method and system based on a sponge city, and the method comprises the steps: responding to an analysis demand and range, and constructing and configuring a water quality monitoring data analysis processing layered architecture; collecting water quality monitoring data and carrying out pretreatment and storage treatment; performing batch processing on the static data, and / or performing streaming processing on the real-time dynamic data to generate a preliminary analysis result; importing the preliminary analysis result into a pre-constructed data model for processing, and iteratively optimizing parameters of the corresponding data model and algorithm parameters deployed by the data model; and actual monitoring data after a deep analysis result output by the application data model is observed and tracked, the architecture is actually optimized and adjusted, and a data result is converted into a visual form, encrypted and provided for an access object. The full-automatic management process provided by the invention not only improves the efficiency of data acquisition and analysis, but also greatly enhances the accuracy and intelligent level of the result, and reduces the supervision cost.

This invention discloses a seedling quality grading monitoring device and method based on image and near-infrared spectral phenotypes. The monitoring device includes: a truss, a camera module, a near-infrared spectral sensor module, a development board, a classification action device, a step-down board, a battery, and an API server. The monitoring method includes: acquiring RGB images of seedlings through the camera; obtaining quantitative data such as leaf size, vein length, and yellow leaf spot area through image processing; acquiring spectral curves through the near-infrared spectral sensor module; extracting quantitative data through signal processing; and then processing and classifying the data, loading it into a classification model for classification, and determining the seedling quality. This invention achieves black-box interface access and hostless operation, reducing device size and cost while improving algorithm speed. Through mobile APP development and web-based host computer development, the hardware device is mountable while reducing learning costs.

This invention relates to a physical layer network coding coordination scheduling method for linear multi-hop underwater acoustic networks. This method introduces PNC (Physical Number Coordination) technology into the link access strategy design of linear multi-hop underwater acoustic networks, overcoming the inherent inter-hop interference limitations of such networks. The method includes the following steps: all sensor nodes transmit local data packets in time slot 1; all sensor nodes except the upstream and downstream nodes simultaneously receive data packets from upstream and downstream neighbor nodes in even-numbered time slots, obtaining coded data packets after PNC mapping; the downstream sensor node receives data packets from upstream neighbor nodes in even-numbered time slots; all sensor nodes transmit data packets received in the previous time slot in odd-numbered time slots; and the sink node completes data packet reception in even-numbered time slots. Compared with existing technologies, this invention exhibits better performance in channel utilization, throughput, and end-to-end latency, especially in channel utilization, which can reach half of the theoretical upper limit.