114results about How to "Reduce space overhead" patented technology

The invention provides a carbonate rock physical parameter seismic inversion method, which belongs to the field of petroleum geophysical exploration. The method comprises the following steps: (1) carrying out pre-stack AVO three-parameter inversion based on a pre-stack angle gather to obtain formation elasticity parameters M, wherein M includes P-wave velocity, S-wave velocity and density; (2) establishing a statistical rock physical model based on logging data; (3) randomly simulating physical conditions of a reservoir based on the rock physical model obtained from step (2) to obtain a random simulation result; and (4) carrying out Bayes classified simulation on the random simulation result obtained from step (3) to obtain a posterior probability distribution, and taking the M obtained from step (1) as the input of inversion to find a R with the maximum posteriori probability distribution, wherein R is the final inversion result.

The present invention provides a map pathfinding method, dividing plane map into one or more identical area, generating waypoint on area border; when finding path, processing path find based on the area to acquire area path; then generating final path based on the triangular facet Path finding according to the acquired area path. The invention also provides a mapping road finding system, adopting method and system of the invention not only saves space occupation, but also improves road finding efficiency, also suitable for process of all kinds of maps.

A technique for monitoring an execution of a program is provided. At least one relational query is constructed and compiled into a program instrumentation. The program instrumentation is incorporated into the program and one or more query results are output from the program as the program executes.

In an MRAM array based on MTJs, the size of segmented word line select transistors and their associated connections become a significant overhead, especially when the operating point is chosen deep along the hard axis of the asteroid curve. This problem has been overcome by placing the big segmented word line select transistors under the MTJ array and reducing the overall MRAM cell array down to a level comparable to a simple Cross Point MRAM array.

The invention provides a method for calculating a fracture description parameter. The method of the invention aims to obtain a fracture development strength body having relatively stable anisotropy medium characteristics based on an early stage pre-stack fracture detection technology, using a Bayes classification method through establishing a rock physic model of the fracture to expand a fracture intensity and a fracture porosity which are obtained through well logging to development bodies of fracture intensity and fracture porosity of a whole working area in order to provide accurate fracture prediction result to comprehensive modeling and spatial distribution of a fracture type reservoir which are carried in a later stage. The method for calculating the fracture description parameter quantitatively overcomes problems that a regular rock physical parameter inversion method is completely based on a statistics hypothesis and performs black box inversion. Besides, the fracture physical parameter inversion is performed on the basis of practical data of a carbonate rock reservoir in western China, and has achieved a satisfied result.

The invention discloses a system for reminding about health schedules. The system comprises: a plan database comprising a plurality of action plans, a user database comprising user information, a plan task processing module, a reminder sending module and a user terminal, wherein the plan database, the user database, the plan task processing module and the remind sending module are arranged in a system server, and the system server is connected with the user terminal through the Internet or wireless communication. The plan task processing module scans plan items in the plan database at regular time to look for a plan item which is matched with the user information in the user database and accords with time conditions, and then sends the plan item to the user terminal through the reminder sending module. Accordingly, the invention also discloses a method for reminding about health schedules, and the method is realized by the reminding system.

The invention relates to a range query method and device based on Trie. The method comprises the following steps that: data preprocessing: coding all position points in a dataset into a character string of which the length is n, and sorting the position points in the dataset according to a character string suffix and a lexicographical order, and numbering ID (Identity); spatial index establishment: according to a sorted character string prefix, constructing a Trie spatial index structure; and range query: on the index structure, obtaining the position points in a query domain, filtering to obtain a candidate set, and verifying and returning all position points in the query range to the candidate set. By use of the method, on the basis of the character string prefix, parts of Trie index structure are constructed, query efficiency is high while space overhead is lowered, a geohas spatial index method is adopted to effectively support the query of any range.

The invention discloses a financial time series similarity query method based on K-chart expression. The method comprises the following steps of feature extraction, index construction and query processing. The method comprises the following concrete steps of firstly, extracting basic mode and classic mode features for a financial time series based on K-chart expression, and respectively translating the basic mode and classic mode features into a basic string and a classic string; secondly, respectively constructing reverse indexes on the basic string and the classic string; for each query sequence, after the basic mode and classic mode features are extracted through the same way, respectively querying the two constructed reverse indexes to acquire two candidate sets, and then carrying out intersection operation to obtain a final candidate set; obtaining a final query result through follow-up processing. The financial time series similarity query method based on K-chart expression can effectively realize nearest neighbor query, has higher measurement precision and query efficiency, has favorable extensibility for time series length, nearest neighbor query scale and data set scale, and can play a significant role in the widened electronic finance trade market.

The invention relates to an index tree construction method for a key value storage system, which comprises the following steps of: sorting and dividing prefixes of key values of key value data to generate a dictionary tree serving as an upper layer structure of an index tree; constructing a hash table by using the hash value of the key value, and generating a lower layer structure of the index tree by using the hash table; and establishing a corresponding relation between the key value data, the hash table and the dictionary tree, and generating the index tree. According to the index tree construction method, key value data indexing operation is conducted through the constructed index tree of the upper-layer and lower-layer structure, more excellent single operation capacity O(L+K) is achieved, lower space cost and higher efficiency are achieved, and range finding and dynamic data increasing processing are supported.

The invention provides a mixed indexing method for metadata attributes. The method comprises the following steps of: dividing the metadata attributes into a high-frequency metadata attribute and a low-frequency metadata attribute according to the used frequency, creation time and recently-accessed time of each metadata attribute, and aiming at the properties of the high-frequency metadata attribute and the low-frequency metadata attribute, establishing indexes by using KD-tree and B-tree trees and an artificial immune algorithm. The invention also provides a structure of a novel metadata management system, and introduces functions and processes of main modules. In the mixed indexing method, aiming at the problems of large time and space expenditure, large exceptional space and the like during the management and searching of metadata, the efficiency of searching the high-frequency metadata attribute is improved, and the space expenditure for managing the low-frequency metadata attribute is reduced.

The invention provides a sparse neural network processor based on a systolic array. The sparse neural network processor comprises a storage unit, a control unit, a sparse matrix operation array, a calculation unit and a confluence array. The storage unit is used for storing weights, gradients, features and instruction sequences used for scheduling data streams. The control unit takes out data required by the training and reasoning process from the storage unit according to the control of the instruction sequence, converts the data into a sparse matrix operation format and sends the data into the sparse matrix operation array. The sparse matrix operation array comprises a plurality of processing units connected in a systolic array mode and is used for completing sparse matrix operation. Thecalculation unit is used for completing element-by-element operation such as a nonlinear activation function. The confluence array delivers the same data segment to different rows of the systolic array through internal data transfer to reduce storage overhead. The processor makes full use of the sparsity of the weight and the characteristics, achieves the improvement of the speed and power consumption ratio in the neural network training and reasoning process, and has the advantages of high concurrency, low bandwidth requirements and the like.

The invention discloses a moving target detection method based on gray features. The method comprises the following steps of: carrying out gray processing and denoising processing on a real-time video monitoring image sequence by utilizing a computer terminal, selecting a multiframe video monitoring image from the image sequence to form a training image sequence, extracting the corresponding gray feature set to each pixel at the same coordinate position in the training image sequence, screening the gray feature sets to obtain a background model, carrying out moving target detection on a subsequent video monitoring image in the image sequence, and carrying out morphology filtering and connected region detection on a binaryzation image including the moving target to obtain an outline and a position of the moving target. The method provided by the invention only needs the searching and comparison operations in the gray feature sets corresponding to the pixels without without complicated parameter estimation or probability calculation, thus simplifying the detection process, improving the detection precision, reducing the time cost and improving the operating efficiency.

The invention discloses a graph structure matching-based social network analysis method, and relates to the fields of a graph structure matching technique, an incremental calculation technique and the like. The graph structure matching-based social network analysis method mainly comprises the steps as follows: (a) a graphic interface helps a user manage graph data, and constructs a mode pattern and a visual query result; and (b) a query engine invokes a distributed algorithm to carry out structure matching calculation; meanwhile, the matching results are evaluated; and top-K results are selected and are visualized on the graphic interface; and (c) incremental calculation is carried out on the existing matching results by an incremental algorithm, so as to ensure the accuracy and the integrity of the results. Aiming at ''large data'' in a social network, efficient and convenient data analysis and maintenance are achieved through distributed, visual and incremental calculation techniques, so that a key technical support is provided for hotspot applications such as expert recommendation, social circle identification and social position analysis.

The present invention provides a generation method and apparatus for index database of points of interest attribute, relates to database field. The invention is invented for resolving the problem of large space consumption of index database in prior technology. The method includes: counting different literal units appeared in the points of interest attribute; obtaining coding of the literal units; searching points of interest attribute containing the literal units and location of the literal units in the points of interest attribute; according to a difference value of the points of interest number corresponding to the points of interest attribute and the location of the literal units in the points of interest attribute to generate index recording list of the literal units; according to length of the index recording list of the literal units to generate address information of the index recording list of the literal units; composing coding of the literal units, index recording list of the literal units, address information of the index recording list of the literal units to generate the index database of the points of interest attribute. The invention can reduce storage space of the index database of the points of interest attribute effectively.

The invention relates to a repeating data and deleted data placement method and device based on load balancing. Based on various kinds of distributed repeating data deletion systems, by virtue of a strategy of changing data block placement, read performance of files is further improved under the condition that a repeated deletion rate is invariable. The method is characterized in that all the contained data blocks are placed by taking single write IO as a basic unit, so that the data blocks in the same IO are independently placed on relatively independent storage nodes as many as possible. Therefore, loading bottleneck during file reading can be eliminated to the greatest degree, parallel maximized utilization of each independent node is realized, and the read performance of the system is improved.

The invention belongs to a method of judging whether a point is in a polygon, pertaining to a computer graphic and computer geometry technical field. The invention first convexly decomposes a polygon and builds a binary tree to manage these convex polygons; then the invention investigates the built binary tree according to the coordinate position of the tested point to find the convex polygon to which a leaf node of the binary tree corresponds; finally the invention tests whether the point is in the convex polygon or not, thus judging whether the tested point is in the given polygon or not. The invention has the advantages that the calculating speed is high and the storage space is saved compared with other similar methods based on polygon convex decomposition. The invention is applied to the condition that a plurality of points are judged whether in the same polygon.

The invention discloses a wear balancing method of a non-volatile memory system based on a multi-level cell. The wear balancing method comprises the following steps: dividing each wafer in a storage region of the non-volatile memory system into a plurality of sub-storage regions; randomly exchanging data in a physical line in the sub-storage regions by using an algebra-based wear balancing algorithm after P writing requests are performed in each sub-storage region every time; and performing data exchange by selecting a hot sub-storage region and a cold sub-storage region in various sub-storage regions after T writing requests are performed in the storage region of the non-volatile memory system every time, wherein T is a region exchange interval, which is a predetermined value or a random number; the hot sub-storage region is the sub-storage region having more cumulative writing numbers; and the cold sub-storage region is the sub-storage region having less cumulative writing numbers. In combination with a table-based wear balancing algorithm and the algebra-based wear balancing algorithm, the wear balancing method disclosed by the invention has the advantages of being long in service life, safe and reliable.

The embodiment of the invention provides a high accuracy division operation method. The high accuracy division operation method comprises setting a system significant bit width 1, searching a start bit position Ps of an unsigned divisor x in a mode starting from a sign bit from high bit to low bit to obtain the significant bit length of the unsigned divisor x, dividing the unsigned divisor x into an a comprising a high 1bit significant bit and a b comprising a residue significant bit and transforming the reciprocal of the unsigned divisor, performing normalization processing on the a, and searching a pre-stored reciprocal table and performing regression processing to obtain the value of 1 / a and further to obtain the value of 1 / (1+b / a). The invention further provides a high accuracy division operation device. The high v division operation method and device improve the accuracy of a quotient by setting a pre-stored reciprocal table or / and through a mode of equation extension, are suitable for scenes requiring relatively high accuracy and are also suitable for scenes where the operational speed is high and the accuracy errors are controlled within a certain range, not only avoid multiple shift operation and subtraction operation in division operation in the prior art, but also avoid over-large errors caused by interpolation fitting.

A RAID-configured grouped-object storage system provides reduced storage space overhead for small objects. The storage system includes a plurality stripes arranged across a plurality of physical objects. Each stripe includes a plurality of storage blocks that are each mapped on to a respectively different physical object. The storage system also includes a plurality of virtual objects each containing at least one storage block. A group of virtual objects is formed when a virtual object contains less storage blocks than the number of stripes by associating the virtual object with at least one virtual object containing less storage blocks than the number of stripes and / or at least one storage block containing zero values so that the storage blocks of each group of virtual objects equals the number of stripes. The storage blocks of each virtual object and of each group of virtual objects are mapped to a respectively different stripe.

The present invention discloses a space-time efficient binary image binary logic operation method based on run-length coding, which can finish the four basic logic operations of or operation, and operation, xor operation and logic difference operation of the binary images and the combination thereof with less space and time expenses. According to the present invention, the binary images participating in operation are expressed into an equivalent run length tree structure by the run-length coding; various position relations between two run lengths are defined, the corresponding run length operation rules are determined according to the different position relations, and the basic logic operations of the binary images corresponding to the run lengths can be finished equivalently; the complicated logic operations are resolved into a plurality of basic binary logic operations, and a run length tree of the operation result binary images is obtained by applying the corresponding run length operation rules; and finally, the result binary images are reconstructed by the run length tree. According to the present invention, by the run-length coding, the space expense expressing the binary images is reduced, and a plurality of series pixel logic operations are compressed into a simple run length operation, thereby reducing the space and time expenses of the method remarkably.

The invention discloses a content addressable memory unit based on memristors and CMOS transistors, reading / writing operation, and a data search matching method therefor. Through the resistive randomand nonvalatile features of the memristor, the content addressable memory unit is formed from two memristors and two CMOS transistors in a special connection mode. A content addressable unit inheritsthe advantages, such as small size, low power consumption and strong expansibility, of the memristors. Compared with a conventional content addressable memory unit, the content addressable memory unitsupplies a larger storage space and achieves simpler reading / writing operation and data search matching operation. The invention provides a new idea for design of the content addressable memory. Thecontent addressable memory unit is composed of the memristors and transistors, so that the reading / writing operation and data search matching operation are achieved by applying different working voltages to input / output lines of the unit; the content addressable memory unit is simple and requires fewer steps in the reading / writing operation.

The invention discloses a subgraph querying method, and belongs to the technical field of computers. The method comprises the following steps: acquiring queried graphs, wherein frontier sets of the queried graphs have a sequential relationship; acquiring current snapshot of graph dataflow on the basis of a time window; acquiring a current extension chain table of the queried graphs on the basis ofthe queried graphs and the snapshot, wherein the extension chain table comprises a plurality of chain table nodes, and L^i in the chain table nodes is used for storing omega (Preq(e_i)); and acquiring a matching solution of the queried graphs according to the extension chain table, wherein the matching solution of the queried graphs meets a time-order matching condition. Time-order characters ofthe queried graphs are considered, the obtained time-order matching between the matching solution and the queried graphs is ensured, and the accuracy of a subgraph querying process is improved.

The invention discloses a method for estimating a small-temperature-difference target distance, which is mainly used for solving the problems that the existing method for estimating the small-temperature-difference target distance is not suitable for an early warning and head-on attacking weapon system and the result is not unique. The method is realized through the following steps of: acquiring a difference signal between target radiation and background radiation on a 3-5mum waveband and an 8-12mum waveband by utilizing a missile-borne dual-band infrared imaging sensor; calculating the temperature difference between a target and a background; establishing a target distance formula R=sum(kiVi) + k5, i being 1 to 4, and calculating variables V1, V2, V3 and V4 in the formula; determining and tabulating factors ki under different weather conditions, wherein i=1-5; looking up the table according to the real-time weather condition to obtain the factor ki; and introducing the variables obtained through calculation and the factor obtained through looking up the table into the target distance formula to calculate the target distance R. The method has the advantages of strong stealth capability, few measurement variables, simplicity in calculation and strong feasibility, and is suitable for estimating the target distance under the background temperature not higher than 200-500 kelvins when the temperature difference between the target and the background is not higher than 5 kelvins.

The invention provides a method for realizing inter-partition communication of an embedded partition operation system. The method is based on lock-free annular channel caching. By means of conditionalcompilation, partition communication is configured, when a user channel cache is configured and used, the channel cache is defined in the user partition space, if a task needs to construct and send amessage, the channel cache operation in the user partition space can be carried out without switching a kernel mode, and one-time copying of the user partition space to the channel cache is omitted.According to the invention, channel buffering based on the annular queue is adopted; the lockless communication of one writer / at least one reader can be realized without relying on hardware; or lock-free communication of a plurality of writers / readers is realized on a hardware platform supporting CAS operation; the real-time performance and safety of the operating system are improved, the verification and test expenditure of the operating system is optimized, meanwhile, two memory optimized conditional compilation configuration options are provided, two times of copying are reduced compared with a traditional implementation mode after starting, and the space expenditure of communication is greatly reduced.

The present invention relates to a weak fingerprint repeated data deletion mechanism based on a flash memory solid-state disk. Oriented to online applications with repeated data, the mechanism implements a deduplication system in a solid-state disk by using hardware resources, and deduplication is performed through weak fingerprint calculation and byte direct comparison, so that the defects of the conventional deduplication technology and pre-hash deduplication technology are overcome, a higher deduplication throughput and deduplication rate are achieved, and overheads on the aspects of time, space, hardware resources and the like are minimized. Moreover, due to a low hardware overhead, parallelism in the solid-state disk can be fully used, and a plurality of deduplication parts perform deduplication in parallel, so that a write delay is greatly reduced. The mechanism is capable of combining with a flash memory solid-state disk properly, is capable of effectively eliminating repeated data, and improves the service life and performance of the solid-state disk.

A method for version upgrade of wireless terminal comprises the follow steps: dividing the version file of the terminal, verifying, generate a terminal version verification value file and saving the file to the wireless terminal; When a version upgrade request is detected, uploading a terminal version check value file to a version upgrade server; dividing and verifying the new version file to generate a new version verification value file; Comparing the new version check value file with the terminal version check value file, generating an upgrade package and sending the upgrade package to thewireless terminal for version upgrading. The new version check value file is saved to the wireless terminal, and the terminal version check value file stored in the wireless terminal is replaced as the terminal version check value file when the next version is upgraded. The invention also provides a wireless terminal version upgrading system, which makes the version upgrading of the wireless terminal no longer rely on a large number of differential packets, and reduces the space overhead occupied by the memory when upgrading.

The invention discloses a wearable device data optimization processing method based on the editing distance. The method includes the steps of firstly, conducting simple zero crossing point data segment on data fed back by a sensor; secondly, conducting periodic detection, and conducting the following more detailed periodic distinguishing if the data conforms to periodic action standards after periodic detection is completed; thirdly, conducting editing distance calculation on a data stream of the second step, and obtaining the calculation result; fourthly, obtaining the final precise result according the result of the third step and through the assistance of a correction function, and making the precise result conform to the objective common sense; fifthly, putting forward an undirected graph concept for a multi-sensor system, and calculating all edge weight values through the similarity algorithm; sixthly, obtaining the minimum sensor set Vf through the optimal node algorithm. According to the method, the existing long data processing flow in a single-sensor system is abandoned, and in a multi-sensor system, the aims of saving energy and reducing space expenses are achieved through an undirected graph and a dynamic planning function.

The invention provides a JPEG image lossless compression and decompression method, system and device, and the method comprises the steps: decoding a JPEG image part, obtaining a DCT coefficient and a quantization table of brightness and chromaticity, and performing caching; sorting the DCT coefficients according to corresponding frequency bands, and dividing the sorted DCT coefficients into a plurality of frequency groups; for the frequency groups, extracting contexts from DCT coefficients of frequency bands to which the contexts belong one by one by adopting a convolutional neural network; and predicting probability distribution of the DCT coefficient according to the context, and performing arithmetic coding to obtain a binary code stream. According to the method, on the premise of only partially decoding the JPEG image, the compression ratio of the JPEG image is improved, and no content in the image is changed. The JPEG image coding format which is widely used at present is compatible.

The invention provides an NAND FLASH garbage collection balance optimization method. Garbage collection related factors such as the effective data volume, the writing frequency and the abrasion degreeof the collection block are comprehensively considered; the NAND FLASH physical blocks are clustered according to the factors, a low-programming-frequency class set, a high-writing-frequency class set and a low-abrasion-degree class set are selected from clustering results to serve as initial screening sets, the intersection of the three initial screening sets is solved, and an optimal recovery block set is obtained. Recovery efficiency and abrasion balance are considered, related factors in the garbage recovery process are considered in a balanced mode, the range of selectable recovery blocks is gradually narrowed through the set intersection idea, and system space expenditure is saved.

Embodiments of the invention provide a dynamic compilation method and apparatus. The method comprises the steps of obtaining a predicted value of a parameter of a to-be-compiled hotspot method from a code cache of a virtual machine; if it is determined that an input value of parameter input is different from the predicted value, updating the predicted value to the input value; replacing a corresponding variable in an intermediate code with the predicted value after converting the to-be-compiled hotspot method into the intermediate code; and optimizing the replaced intermediate code and converting the optimized intermediate code into a to-be-executed local target code. According to the dynamic compilation method and apparatus provided by the embodiments of the invention, the scale of codes generated by compilation can be reduced and the system performance of the virtual machine can be improved.