68results about How to "Reduce omission rate" patented technology

A multiple pornographic image classification method based on an image segmentation algorithm and deep learning relates to the technical field of information, especially to the technical field of image identification. The method is characterized by comprising the following four major steps: skin color identification, principal component analysis of the skin color region, deep learning, and pornographic image identification based on a convolutional neural network. First, non-pornographic images are screened out through a skin color pixel detection and skin region partitioning algorithm based on the YCbCr theory. Undetermined images are input to a LeNet5-based convolutional neural network model for identification after feature extraction. Compared with the traditional identification based on skin color and features, the method can eliminate the noise influence of non-body-part images, is not constrained by light and human postures, and can greatly improve the accuracy of traditional pornographic image classification. Compared with general deep learning based on a convolutional neural network, the method does not need massive labeled images, and the characteristics of a deep residual network determine that the model can better analyze the characteristics of pornographic images. Only through about ten hours of training, an identification effect above 90% can be achieved.

The invention provides a completeness checking method and device for sub-table data. The completeness checking method and device are applied to a distributed server cluster. The device comprises a control server and a business server. Business data in the business server is split and stored in a plurality of different data sub-tables. The method comprises the steps that the control server collects the business data in all the data sub-tables into a local preset data table; before a timing processing task against the preset data table is triggered, a data checking request for all the data sub-tables is sent to the business server in a timing mode; the business server executes data completeness checking on all the data sub-tables responding to the data checking request and supplements and sends business data to the control server; and the control server supplements the business data collected in the preset data table. Through the completeness checking method and device, the completeness of the business data in all the data sub-tables when the timing task is triggered can be ensured to the maximum extent, and the probability of omission of to-be-processed business data is effectively lowered.

The invention provides a vehicle video recognition method. The vehicle video recognition method comprises the following steps of decoding video data of a monitoring device and separating data frames to form image data of each frame of videos; performing vehicle license plate recognition to each frame of images and building a video image vehicle license plate recognition coinciding degree list; calculating probabilistic list data of a vehicle license plate recognized from continuous adjacent frames of images and locking vehicle license plate information in videos; calibrating vehicle license plate recognition accuracy again according to the vehicle license plate recognized in the videos of the monitoring device and accordingly acquiring accurate vehicle recognition effect. The vehicle video recognition method aims at fully applying the existing monitoring device, further finishing perspective mastering of monitored video information, improving effective data utilization of public security department, security and protection department, traffic department, traffic police department and the like, improving automation degree and resource utilization rate, building an intelligentized monitoring application platform for recognition, analysis, application and aid decision making and with the monitoring device as an intelligentized node.

The invention relates to an intelligent processing method for infrared inspection data of a power transmission line. The method comprises the following steps: S1, inputting an infrared inspection picture; S2, renaming an input picture; S3, performing component identification and regional temperature detection analysis on a renamed picture; S4, manually auditing defects; S5, generating defect information by using a defect generator; and S6, recording defects by using an RPA robot and notifying team managers. According to the invention, infrared inspection photos are automatically analyzed, whether related parts emit heat is detected, and the defects are automatically recorded into the system, so that a lot of manpower and material resources can be saved, the workload of personnel is reduced, the working time is shortened, and the working efficiency is improved; the pertinence and accuracy of intelligent processing of the infrared inspection data of the power transmission line are improved through renaming; the infrared inspection data processing efficiency of the power transmission line can be effectively improved.

The invention provides a vehicular video monitoring method for vehicular access entries. The method includes the steps of erecting a monitor camera at a vehicular access entry; allowing the monitor camera to monitor entering vehicles by video; forming image data of each video frame by video data of the monitor camera; establishing a list of video image license plate recognition fitness for vehicular license plates for each image frame; locating the license plates in the area of an image according to recognized results of the license plates in the video of the monitor camera; and calculating data of a list of probability of recognizing the license plates in continuous adjacent image frames, and locking the information of the license plates of accessing vehicles in the video. The vehicular access entries are subjected to monitoring video acquisition, monitoring video is applied and analyzed, a database is comprehensively processed, perspective mastering of vehicular information is completed, and accordingly data of park vehicles and users are utilized effectively, automation level and resource utilization rate are increased, and an intelligent park application platform using monitor equipment as an intelligent node is established.

The invention relates to a warehousing logistics system based on UHF RFID technology. The warehousing logistics system comprises RFID tags, an RFID reader-writer, RFID antennas, a single-chip microcomputer, a photoelectric switch, a background server and a display screen, wherein the RFID tags are embedded on the surfaces of cargo boxes, the RFID antenna is connected with the RFID reader-writer, and the RFID antennas are symmetrically arranged on vertical door frames on both sides of a logistics gate, the RFID reader-writer is connected to the single-chip microcomputer through a serial port, the single-chip microcomputer is connected with the photoelectric switch through a GPIO interface, the photoelectric switch is arranged under the RFID antennas, the single-chip microcomputer is connected to the background server through the network, the background server is connected with the display screen, and the display screen is arranged on the wall at one side of the logistics gate. The warehousing logistics system can carry out the batch scanning of RFID tags and complete the scanning at the same time as goods entering and leaving a warehouse. When the goods normally pass through the logistics gate, the information of all the goods can be collected in real time without waiting.

The invention discloses a face recognition method based on a video stream. The method comprises the steps that multi-frame continuous contrast recognition for high-definition videos is adopted to decode video data of monitoring devices, and data frames are separated to form image data of each frame video, so that the face recognition rate is substantially increased in an exponential order; face feature recognition processing is performed on each frame image to establish a probabilistic arrangement list of face recognition in each frame image; through calculation of fit-goodness list data of face recognition in continuously adjacent frame images, face information in the videos is locked, and then simultaneous recognition analysis is performed on faces of multiple targets to lower the omission rate of the face information; and according to the faces recognized in the video of each monitoring device, face recognition accuracy is calibrated again, so that a more accurate recognition effectis obtained. According to the method, through sufficient application of the monitoring devices, the big data control capability of a city administrative department based video monitoring is improved,and the automation degree is increased.

The invention provides a traffic flow monitoring system based on an automobile electronic marker and a monitoring method thereof and relates to the field of traffic flow counting. The traffic flow monitoring system comprises a console, a display device, a marker reader and an electronic marker; the marker reader arranged beside of a lane of an expressway is in wireless communication with the electronic marker arranged on an automotive vehicle and is used for transmitting communication data to the console in real time; the console is used for processing the communication data and displaying the communication data through the display device; the traffic flow monitoring system further comprises a pressure sensor embedded on the ground of the lane; the pressure sensor is used for transmitting a collected pressure value to the console in real time; and the console is used for judging whether the marker reader at a dormant state is awakened or not according to the pressure value. According to the traffic flow monitoring system provided by the invention, the electronic marker is awakened through utilizing pressure on the ground in an automobile driving process, and the marker reader in wireless communication with the electronic marker, so that a low omission rate is realized; and a communication behavior consumes a lot of power and a dormant-awaking mechanism is added, so that a radio frequency signal is transmitted by the marker reader only if an automobile passes through, the waste of electric energy is avoided and high-efficiency energy saving is realized.

The invention discloses an operation sequence selecting method suitable for an STP operation test. The operation sequence selecting method suitable for the STP operation test comprises the steps of abstracting a whole railway yard into binary tree representations of different tree roots according to throats by using station tracks as boundaries, yard edge segments as root nodes, station tracks asleaf nodes and equipment on a path as other nodes of a binary tree; carrying out depth-first search on each tree and intensively marking passing segments; and after all searches are finished, using marked information to calculate an occupation error and a clearance error of each segment. According to the operation sequence selecting method suitable for the STP operation test, through tree form modeling, depth-first search and undetected segment statistics, the coverage of operation segments is guaranteed, and the omission probability of completely depending on manually following a vehicle to review the lengths of segments is effectively reduced. Through statistics of operation conditions and results of different segments, statistical analysis of the segment length error is automatically output so that the working intensity of manually following and watching the vehicle is effectively reduced, and the statistical accuracy is improved.

The invention relates to a seed watermelon harvesting collection machine. The seed watermelon harvesting collection machine is characterized by being composed of a rack, a circular-arc-shaped seed watermelon separation device, a weed raking mechanism, a seed watermelon stopping plate, a seed watermelon stopping plate reinforcing rib, a reinforcing rib fixing plate, a seed watermelon seedling separation device, a spiral conveying device, a combined chain wheel and a depth limiting device. According to the seed watermelon harvesting collection machine, seed watermelons are separated through the circular-arc-shaped seed watermelon separation device at first; then, the seed watermelon seedling separation device is used for separating seed watermelon seedlings; a cutter of the weed raking mechanism is used for cutting off weed and the seed watermelon seedlings, which are wound on a propelling blade and a blade mandrel; and finally, the seed watermelons are separated and gathered through the spiral conveying device. Totally, the invention provides the seed watermelon harvesting collection machine which can be used for reducing the weed wound on the seed watermelon seedlings, reducing a fault rate of machine tools, improving the working efficiency and also reducing a damage rate of the seed watermelons and the like; seed taking work is convenient to carry out and parts are convenient to detach; and the operation of the whole machine is convenient and reliable.

The invention discloses an intelligent postbox cabinet, and relates to the technical field of postbox cabinets. When a user takes away a letter, it is inconvenient to take the letter because of poor sight, and when there are many letters, there may be a probability that a letter will be missed in a letter receiving groove. The intelligent postbox cabinet of the scheme comprises a cabinet body, wherein a side wall of the cabinet body is provided with a plurality of letter receiving grooves, and notches of the letter receiving grooves are rotationally connected with cover doors; a lower slot wall of the letter receiving groove is provided with a driving slot, and a side edge of the cover door near a rotating shaft is fixed with a driving strip; a letter receiving box is slidably arranged inthe letter receiving groove, a first elastic piece is connected between the letter receiving box and the groove bottom of the letter receiving groove, a locating slot is opened at the bottom of the letter receiving box, and a locating bar for turning into the locating slot and fixing the letter receiving box is arranged on the lower groove wall of the letter receiving groove; and a linkage rod group is connected between the locating bar and the driving strip. When the user opens the cover door, the letter receiving box can pop up, so that the user can more conveniently take away letters and reduce the probability of missing a letter.

The invention relates to the technical field of fresh agricultural product distribution and provides a moveable type automatic dispatching and sorting electronic weigher and an automatic dispatching and sorting weighing method thereof. The moveable type automatic dispatching and sorting electronic weigher comprises an instrument, a code scanning device, a connecting rod and a weigher body. One end of the connecting rod is connected with the instrument, and the other end of the connecting rod is connected with the weigher body. The code scanning device is in wireless connection with the instrument. The instrument further comprises a master control unit, a communication module, a display module, an input module, interfaces, a printer and a wireless transceiver. The master control unit is connected with the communication module, the display module, the input module, the interfaces, the printer and the wireless transceiver. The weigher body further comprises a weighing base, a sensor, a power module and universal wheels. The power module is arranged inside the weighing base. The universal wheels are installed at the bottom of the weighing base. The sensor is connected with the instrument through the sensor interface. The moveable type automatic dispatching and sorting electronic weigher and the automatic dispatching and sorting weighing method thereof can be used for automatic sorting reading, statistics and completion of sorting records, greatly reduce manual operation, are low in error rate, and are more efficient and rapider.

The invention provides a three-dimensional model interest point extraction method and system based on hierarchical learning. The three-dimensional model interest point extraction method comprises a process of training a layering device for interest points of a three-dimensional model and a detail point extractor , and a process of performing interest point prediction on the three-dimensional modelto be detected, wherein the process of training the layering device of interest points of the three-dimensional model and the detail point extractor comprises the following steps: for any one three-dimensional model in a set D, dividing all manually marked interest points into common interest points and detail interest points; extracting feature descriptors of all vertexes of the surface of the three-dimensional model by utilizing the multi-feature descriptors; assigning values to the interest points and the labels of the points near the interest points by using an activation function so as to train two neural networks respectively; and combining the two trained neural networks through feature vector matching to obtain a layering device and a detail point extractor which can predict interest points of the three-dimensional model. For the three-dimensional model interest point extraction method, the interest point extraction result is obviously superior to that of a traditional algorithm, and the interest point omission rate and repetition rate are low.

The invention discloses a park comprehensive energy abnormal data combined interpolation method, and the method comprises the steps: taking comprehensive energy data in a park as a sample data set, and obtaining abnormal data in the sample data set; interpolating the sample data set based on the position of the abnormal data to obtain a corresponding interpolation sequence; performing error analysis on the interpolation sequence, and determining an interpolation weight of the sample data set; and based on the interpolation weight, performing interpolation on the user data similar to the characteristics of the sample data set. According to the embodiment of the invention, the data interpolation efficiency of the integrated energy system is improved, and the high precision of the interpolation work is also considered.

The invention relates to a hardness detection device, in particular to an electronic component hardness detection device. For solving the technical problem, the electronic component hardness detectiondevice which can replace manual work and is convenient to use and low in omission rate is provided. According to the technical scheme, the electronic component hardness detection device comprises a base, a first supporting rod, a second supporting rod, a third supporting rod, a cushion block, a conveying wheel, a first belt wheel, a first rotating shaft, a first bearing and the like; the first supporting rod, a supporting table and the second supporting rod are sequentially installed at the upper part of the base from left to right; and the third supporting rod is installed at the upper partof the first supporting rod. The device achieves the effects that the manual work can be replaced, the use is convenient, and the omission rate is low; the device can be used to rapidly complete the hardness detection work of a large number of electronic components, so that the detection efficiency is high, the practicability is high, the working load of workers is reduced, the product quality isimproved, and higher economic benefits are provided for factories; and besides, the device is long in service life.

The invention provides a cell lysis device and a cell purification and/or cell lysis method based on the cell lysis device, and belongs to the technical field of cell lysis purification. The cell lysis device comprises a pulse power supply (1) and a microcapillary electrode (12) electrically connected with the pulse power supply (1), and the microcapillary electrode (12) comprises an electrode tube (1-1); and a liquid metal thin film is attached to the inner wall of the electrode tube (1-1). According to the invention, the pulse power supply (1) is controlled to release a voltage, and a non-uniform electric field is formed above impurity cells through the electrode tube (1-1). As the surrounding voltage is far higher than the transmembrane potential of the impurity cells, the impurity cells are rapidly lysed so as to fulfill the aim of lysing the impurity cells.

The invention discloses an auxiliary inquiry system and method. The auxiliary inquiry system comprises a medical record inquiry content pushing module, an inquiry form generating module and an inquiry module, the medical record inquiry content pushing module is used for receiving previous diagnosis and treatment data, obtained in real time, of a patient, pushing needed feature items according to an embedded knowledge graph and association rules, combining and labeling the features based on a classifier, and providing the features to an inquiry form module; the inquiry form generation module is used for generating an inquiry form and an inquiry sequence according to the tag category and each feature; and the inquiry module is used for obtaining inquiry records after carrying out inquiry according to the inquiry sequence and the corresponding inquiry form, and generating a complete inquiry medical record after filling the inquiry records into corresponding positions of the inquiry form. Therefore, the system provided by the embodiment of the invention can analyze the previous diagnosis and treatment data of the patient, obtain the characteristics related to the inquiry and perform category classification and inquiry sequence determination on the characteristics related to the inquiry, so unified inquiry and unified filling of inquiry records are realized; the inquiry medical record is obtained after unified processing of the medical record data, the inquiry integrity is improved, and the inquiry omission rate is reduced.

Disclosed is a telfon paint coating method. The telfon paint coating method includes that telfon paint made from a same material is coated in different coatings. Due to the fact that a plurality of coatings are generated on the surface of a workpiece, and each coating is thin, and difference and rate of change for the surface of the workpiece coated by a single coating are low, so that flatness of the surface of the workpiece after integral coating can be improved; due to the fact that paint contrasted in colors is adopted to adjacent coatings, a coating effect can be directly obtained by observing colors of the workpiece during coating, missed points of the coatings can be obviously displayed to be found and recognized, and missing rate for coating is lowered.

The invention discloses a method and device for high-purity sorting of rare cells. The method comprises the following steps of (1) spreading a cell sample; (2) carrying out scanning imaging on the spread cell sample; and (3) according to an imaging result, reading position information of target cell and distance information between the target cell and other surrounding non-target cells, and judging the obtained distance information: if the distance information does not meet a set sampling distance requirement, sampling an area where the target cell is located according to the position information of the target cell, obtaining an intermediate cell sample containing the target cell and other surrounding non-target cells, spreading the obtained intermediate cell sample again, and returning to the step (2); and if the distance information meets the set sampling distance requirement, directly sampling the target cell, so that sorting of the target cell is achieved. The method disclosed by the invention is simple and convenient to operate, and has the advantages of high cell spreading speed, high flux, high target cell sorting efficiency, high purity, low omission rate, repeatability and the like.