17results about How to "Guaranteed complexity" patented technology

The invention relates to the technical field of image processing, and provides an image classification method, comprising: acquiring an image to be classified and a classification network, the classification network comprising a first module, a second module, a third module and a fourth module; the to-be-classified image after preprocessing is input to the first module to output a first feature map, the first unit is used for extracting features, and the second unit is used for enhancing feature representation; inputting the first feature map into a second module to output a second feature map; inputting the second feature map into a third module to output a third feature map; and inputting the third feature map into a fourth module to output a classification result. According to the method, by increasing the complexity of feature extraction and enhancement, the risk of overfitting can be reduced while the complexity of a classification network is kept, and more global information can be captured by increasing deeper feature extraction, so that the limitation of local features is reduced, and the accuracy and efficiency of a classification task can be improved.

The invention, which relates to the technical field of the safe transmission of the ship information, discloses an information transmission method for a shipborne Internet of Things terminal, comprising the following steps: collecting the on-site hydrological data in the operation process of a target ship through hydrological collection data, and converting the hydrological data into corresponding hydrological processing data; a hydrological transmission unit converts real-time hydrological processing data to generate corresponding hydrological protection data, a far-end observation platform receives, restores and stores the hydrological protection data, in the conversion process, a corresponding partition list is delimited based on a decimal system of segmented word strings in a first intercepted partition list, and association mapping is established; according to the method, the conversion logic is used for replacing a secret key to convert the on-site hydrological data, the situation of depending on the secret key is avoided, the unnecessary maintenance and management cost is avoided, meanwhile, the far-end observation platform can grasp the situation of the target ship in the operation process in real time, and assistance is further provided for the operation process of the target ship.

The invention discloses a somatosensory dance game method and device for a user-defined dance video and a computer readable storage medium. The method comprises the steps that keywords input by a user and used for the user-defined dance video are obtained; determining the dance level of the user; generating an accompaniment demand according to the keyword and the dance level; according to the accompaniment demand, generating an accompaniment meeting the accompaniment demand through a preset music model; according to the dance level, screening an adaptive initial dance action from a preset dance action library; according to the accompaniment and the initial dance movement, generating a dance movement sequence matched with the accompaniment; generating a lyric demand according to the accompaniment and the keyword; according to the lyric demand, generating lyrics meeting the lyric demand through a preset natural language model; and combining the accompaniment, the dance action sequence and the lyrics to generate a self-defined dance video. The method has the advantages of being capable of meeting personalized requirements of users and improving game experience of the users.

The invention relates to the technical field of security authentication, and discloses an information security and data security authentication and evaluation system and method, and the method comprises the steps: a user sets a password through a password setting module, and judges the security of the password; judging whether the initial password is the same as the password in the existing password set or not through a first strategy, and if yes, prompting the user to reset the password; if not, the similarity of the passwords is judged through a second strategy, and the security of the passwords is further ensured; the identity verification module is used for judging whether the user has the right to access information and data in the database or not, and judging whether a password input by the user is matched with or similar to a password stored in the database or not by adopting a third strategy; according to the method, the levels of password management and data security protection are effectively improved, and the security in the information storage and access process is ensured.

The utility model discloses a damping product manufacturing die, and particularly relates to the field of manufacturing dies, the damping product manufacturing die comprises a bottom plate, material returning rubber is installed on the top of the bottom plate, and a material returning positioning plate is installed on the top of the bottom plate in a matched mode; the device is provided with a rapid assembling and positioning device, butyl products used by a brake system can be formed in a complex mode and have good dimensional accuracy, the consistency and the yield of the products are guaranteed in the manufacturing process, the percent of pass of parts is high, the boundary dimension accuracy is high, and meanwhile when the mold filler strip and the bottom plate are assembled, the production efficiency is high. The aim of conveniently assembling and combining the bottom plate and the mold filler strip is fulfilled; the device is provided with the limiting device, the corresponding face of the movable connecting pad is attached to the outer side of the mold filler strip, the limiting and supporting effects on the outer side of the mold filler strip can be achieved, the movable threaded rod on one side of the movable connecting pad has the guiding effect, the positioning nut on the outer side of the movable threaded rod has the limiting effect, and the service life of equipment is prolonged.

The invention relates to the technical field of optical signal processing, in particular to a depolarization crosstalk method based on a single pilot frequency and a DSCM system. Polarization decrosstalk is realized through a single pilot frequency, so that a subsequent equalizer for equalization of each subcarrier is simplified into two multi-tap SISO equalizers with independent polarization, and the calculation complexity of the equalizer at a receiving end is remarkably reduced. Meanwhile, the protection frequency band between adjacent subcarriers is eliminated, and the spectrum efficiency of the signal is improved. According to the method, the influence of the XY time delay of the receiving end on the low-complexity polarization decrosstalk scheme based on the pilot frequency is considered, and the effectiveness of the polarization decrosstalk method based on the pilot frequency is further ensured by estimating and calibrating the XY time delay of the receiving end.

The invention discloses a parking management method and device based on a block chain technology and a medium, and relates to the technical field of parking management. Vehicle information of a driving vehicle pre-entering a target parking lot is collected and authenticated, and a corresponding reservation number is extracted after authentication is passed; performing protection uploading on the parking record information of the driving vehicle according to the reservation number, determining a first position array and a second position array of the current parking of the driving vehicle based on the randomly generated reservation number in the protection uploading process, and combining the first position array and the second position array with a plurality of groups of processing arrays converted from the parking record information to obtain a slice reconstruction array of each group of processing arrays; after being generated, the slice reconstruction array of each group of processing array depends on the updating of the first position array and the second position array by the previous group of processing array, and the generation of the slice reconstruction array of the first group of processing array depends on the reservation number, so that the dynamic variability of each parking record information is ensured; in this way, the high safety of the parking record information of the running vehicle is ensured.

The present invention belongs to the technical field of sewage treatment devices, and specifically relates to an integrated sewage treatment device based on deep learning, including a detection chamber, a reagent kit, an electrode chamber, a symbiotic chamber, and a filter chamber. The reagent kit is arranged in the detection chamber, the electrode chamber is connected to the symbiotic chamber, the detection chamber is connected to the electrode chamber through a pipeline, the symbiotic chamber is connected to the filter chamber through a pipeline, and the filter chamber is connected to the electrode chamber through a pipeline. The present invention divides pollutants into large groups and small groups according to the treatment methods and multi-dimensional conditions of different pollutants, and finds the optimal current density, pH concentration, H concentration, and pH concentration of the corresponding treatment group. 2 O 2 Concentration and PMS concentration, through the kit and variable power supply to adjust the voltage of traditional chemical electrodes and biochemical electrodes, each group of pollutants in the sewage can be accurately and efficiently treated.

The invention provides an agricultural drought prediction method and system fused with spatial-temporal feature learning. The method comprises the following steps: collecting and preprocessing data; constructing a drought prediction model which comprises a spatial feature extraction module, a spatial-temporal feature learning module and a sequence prediction module; constructing a drought event identification and prediction model; and the drought event identification and prediction model analyzes the standardized soil humidity index SMI sequence of the future continuous M days predicted by the sequence prediction module, extracts the drought event, analyzes and obtains the spatial centroid movement characteristics of the serious drought event, and predicts and obtains the time-space change condition of the future drought. According to the method, time and space attributes of drought are comprehensively considered, high-precision grid meteorological data are utilized, a plurality of hydro meteorological elements are input into the trained convolutional neural network and deep learning coupling model to predict future drought indexes, the complexity of the model is kept within a proper range, meanwhile, early warning is efficiently carried out on the drought, and the drought early warning efficiency is improved. And technical guidance is provided for drought control and water resource management.

The invention relates to a telescopic landing leg type omnidirectional AMR for a narrow roadway, which is characterized in that a wheel train is arranged in a manner of combining a horizontal steering driving wheel with an auxiliary universal supporting wheel, and the omnidirectional AMR comprises the universal supporting wheel at the front part in the advancing direction and the steering driving wheel at the rear part in the advancing direction; the universal supporting wheels are distributed on the two sides of a center shaft of the vehicle body in pairs through first cross-shaped commutators, and the steering driving wheels are distributed on the two sides of the center shaft of the vehicle body in pairs through second cross-shaped commutators. The two universal supporting wheels and the two steering driving wheels are connected with each other through a telescopic device, so that internal and external stretching of the wheels is realized; lifting mechanisms are arranged at the two ends of a middle shaft of the vehicle body. According to the telescopic supporting leg type heavy-load AMR suitable for the narrow roadway scene, the vehicle body height of the AMR is reduced under the condition that the large load of the AMR is guaranteed, so that the gravity center height of carried goods is reduced, and the overturning risk is reduced.

The application belongs to the technical field of special robots, and particularly relates to a high-integration anti-radiation main-backup redundant collaborative robot joint module, which solves the problem that a collaborative robot joint module cannot normally serve due to various radiation effects caused by ray ionization in a nuclear radiation environment; the joint module is composed of a module integrated by a mechanical structure and a redundant comprehensive electronic circuit system; the redundant comprehensive electronic circuit system comprises a low-voltage power supply system, a servo driving system, a communication system, a sensing and perceiving system, an MCU control system and a constant temperature system, and all of them are integrated with main-backup function modules; the main-backup function modules work independently; the MCU control system detects fault signals of the main-backup modules of each system, and controls automatic switching of the main-backup function modules in each system; the redundant comprehensive electronic circuit system of the application is highly integrated and mechanically and electrically assembled with the mechanical structure inside a joint module shell; the shell is made of radiation-resistant lightweight material, and each system inside the joint has high-integration redundancy, radiation resistance and independence.

The invention relates to a cable current-carrying unbalance degree calculation method and device and computer equipment. The method comprises the following steps: acquiring electrical parameters, laying environment data and cable structure data of parallel cables in the same direction, and constructing an environment impedance matrix of the parallel cables in the same direction; constructing an environment cable section model of the same-direction parallel cable based on the data, and identifying each associated parameter of the same-direction parallel cable; constructing a full-line equivalent circuit model of the same-direction parallel cables based on the correlation parameters; and on the basis of the full-line equivalent circuit model, identifying current-carrying distribution information of the same-direction parallel cables, and on the basis of the current-carrying distribution information, calculating the current-carrying unbalance degree of the same-direction parallel cables. By adopting the method, the safety and the operation efficiency of the power system can be improved.

The invention belongs to the technical field of scheduling algorithms, and particularly relates to a heterogeneous computing list scheduling algorithm based on an improved prediction cost matrix. The BPPTS algorithm in the invention comprises two key stages. Firstly, in the task priority ranking stage, the average calculation cost of tasks is considered, and the calculation requirement of each task is efficiently evaluated so as to determine the ranking of task priorities. Secondly, in a processor selection stage, the algorithm considers the out-degree of each task, namely the number of subsequent tasks, optimizes the mapping relation between the tasks and the processors, reduces the waiting time of the subsequent tasks, avoids the problem that the processors are overloaded or idle due to the fact that the tasks with high calculation requirements are distributed to the low-performance processors, and improves the calculation efficiency. And the load balancing and scheduling efficiency is effectively improved.

The invention discloses a high-strength password creating and memorizing method, and belongs to the technical field of high-strength password creating and memorizing. According to the high-strength password creating and memorizing method, association can be carried out according to specific platform functions by adopting the creation of a united password, so that the password is set, various letters, numbers and symbols can be memorized quickly in the same manner, and the memorizing manner is very simple, convenient and fast; the password setting complexity can be completely ensured, the memory of people cannot be affected due to excessive complexity, secondly, a hierarchical management mode is adopted for the passwords of various platforms, the passwords are divided into core passwords, important passwords and temporary passwords, the risk that people use the same password in a large number, and consequently collective theft is caused is reduced, and the security of the users is improved. And the password can be regularly detected by using a regularly used detection tool, so that the condition that the password is stolen and cannot be known by a user is reduced, and the password can be recorded in word or excel for encryption protection.

This invention discloses a rotating metal blade comb structure for a vacuum cleaner floor brush, comprising: a hair cutting mechanism including a metal blade connecting rod and metal blade comb teeth, the metal blade comb teeth and the roller assembly rotating in opposite directions; a hair gripping mechanism, located upstream of the roller assembly relative to the hair cutting mechanism, including a first comb plate and a gripper assembly, a guide assembly, and a linkage assembly arranged sequentially along the surface of the first comb plate towards the roller assembly, the gripper assembly being slidably mounted on the first comb plate, and a pressing block being provided on the linkage assembly towards the guide assembly, the pressing block being associated with the gripper assembly; and a hair adsorption mechanism including a second comb plate and an adsorption element disposed on the second comb plate. This invention can solve the problems of difficult hair removal from common floor brush rollers, where some combing and cutting mechanisms cannot guarantee the efficiency and extent of hair removal during cleaning, and may also damage the brush assembly, affecting the use of the roller brush.

A small target detection method, system, device and medium based on a local attention feature correlation mechanism, the method comprising: data preprocessing, model construction, model training, small target detection; the system, device and medium are used to realize a small target detection method based on a local attention feature correlation mechanism; the application extracts spatial features and semantic features through a residual network, performs feature fusion through a feature pyramid mechanism, reconstructs features based on a local attention feature correlation mechanism, trains a model using a Hungarian matching algorithm, obtains an optimal matching model suitable for a data set, can ensure high accuracy with fewer iteration times, and has the characteristics of low network operation complexity, low training cost, easy model migration, and high model detection performance.

The application provides a photoresist post-baking simulation method, a parameter calibration method, an adjustment method, equipment, a medium and a product. The three-dimensional light intensity distribution in the photoresist is obtained, the initial concentration distribution of the substance in the photoresist after exposure is calculated based on the three-dimensional light intensity distribution, and the initial concentration distribution is processed by using an implicit-explicit hybrid algorithm to obtain the concentration of the developing inhibitor in the photoresist. Therefore, the diffusion term and the reaction term are separated and processed by using the implicit-explicit hybrid algorithm, the overall calculation efficiency is improved while ensuring the diffusion calculation precision, and the photoresist post-baking simulation is more efficient and accurate.