79 results about "Bidirectional search" patented technology

The invention provides a path planning method and system, wherein the method comprises the following steps: mapping roads into nodes in a traffic network topological graph and mapping endpoints of the roads into arc sections in the traffic network topological graph; searching bi-directionally from an original node and a destination node, thus obtaining a plurality of current nodes; calculating the cost from a plurality of current nodes to the original node and the destination node sequentially and obtaining a current node of the minimum cost from the original node to the destination node; and according to the current node of the minimum cost, the original node and the destination node, obtaining a path plan from the original node to the destination node. The method and the system overcome the defect in the prior art that a Dijkstra method of adjacent matrices is adopted to store traffic network topological data, although whether (i; j) is a network edge can be queried within the time O (1), the complexities for querying relevant nodes which are the most significant for the shortest path plan are all O (n), which causes higher query complexity.

The invention relates to the artificial intelligence technology field and particularly relates to an image text matching model training method, a bidirectional search method and a related device. Thetraining method comprises steps that global and local representations of an image sample and a text sample are extracted, a pre-constructed matching model is trained, the global and local representations of the image sample and the text sample are respectively mapped to the specified semantic space by the matching model, and similarity of the global representation and similarity of the local representation are calculated; according to the default weight of the similarity of the global representation and the default weight of the similarity of the local representation, the image and text matching degree is determined through employing a weighted summation mode. The training method is advantaged in that the acquired matching degree is based on considering both detail characteristics of an image and global characteristics, and the acquired matching degree is more accurate and comprehensive.

The invention discloses a UAV path optimization method based on an improved bidirectional rapidly-exploring random-tree algorithm. The method comprises the following steps: 1, establishing a UAV kinematic constraint model and determining a starting point and a target point in a two-dimensional space; 2, based on a backward sector region sampling method, exploring the nodes of the bidirectional random tree; 3, based on a dynamic adaptive step size method, solving the growth limitations near an obstacle; and 4, by connecting the nodes of a bidirectional search tree, forming a complete flight path. Based on a classical B-RRT* algorithm, the method, based on the backward sector region sampling method, restricts random tree sampling points to the backward sector region when the nodes are explored, solves the growth limitations near the obstacle based on the dynamic adaptive step size method, and finally connects respective nodes to generate a flight path that conforms to the real flight ofthe UAV, thereby providing a feasible method for autonomous navigation of the UAV at low altitude.

The invention discloses a quick depth video coding method. In the method, all viewpoints in a multi-viewpoint depth video predictive coding structure are divided into main viewpoints, first-stage auxiliary viewpoints and second-stage auxiliary viewpoints, different quick coding strategies can be adopted according to coding frames with different types in different viewpoints, and the relevance among the viewpoints of a depth video signal is determined by utilizing the information of the coded frames so as to determine whether the current coding frame is subjected to prediction among the viewpoints or not; and whether to perform bidirectional search or not during the coding of the current coding macroblock is determined by utilizing a motion vector of a coded adjacent block or the search mode of the optimal matching block of the current coding macroblock in the current coding frame; and thus, the time-consuming and poor-effect searches in the process of coding the current coding frame is prevented, and the calculation complexity of the multi-viewpoint depth video codes is reduced effectively while the accuracy and compression efficiency of a depth map are ensured.

The invention discloses an AUV mobile data collecting method in an underwater sensing network based on data prediction, comprising that AUV uses a centralized algorithm to initialize a maximum neighbor node density cluster, uses SVR to fit and predict the acquired data and establish a prediction model, and updates the cluster according to the trend similarity degree of the predicted model; the AUVand the corresponding cluster save the same prediction model synchronously; in a data collection process, the access collection of the clusters is skipped and the prediction model is directly used for prediction; when the prediction model of the corresponding cluster is larger than a preset maximum tolerance error threshold or delay sensitive threshold, a update request is sent immediately; bidirectional search is used in order to find the current position of the AUV and a change is notified; and the AUV re-plans the remaining path to obtain a new prediction model. Through data prediction, the AUV traversal path length is reduced to reduce the energy consumption and collection delay of the AUV, the amount of data size of the entire network is reduced, the energy consumption of some nodesis reduced, and the overall network performance is improved.

The invention discloses a MapReduce-based big picture distance connection query method. The method comprises the steps of: (1) extracting initialized query parameters including original picture, accessed nodes, extension range and query result; (2) conducting two-way extension on the original picture on hadoop, starting extension from a source node set and a target node set, conducting each extension on the basis of a cost model, conducting pruning operation by adopting dynamic threshold, and adding newly-extended node to the accessed node set; (3) continuously traversing the incompletely extended remaining nodes until all the nodes meeting the extension range are extended; and (4) after completing iteration, recording route query results between the target nodes and the source nodes in the accessed node set, and returning a query result. The invention provides a cost model-based self-adaption method under the MapReduce environment, reduces the extension space and iteration times by utilizing dynamic threshold-based pruning two-way searching algorithm and Segment index, and improves the task execution efficiency.

A method for traversing a multiplexed data packet stream including choosing a starting location in a multiplexed data packet stream, making a bi-directional search from the starting location to locate a nearest system stream object; and traversing the system stream as a doubly-linked list of objects. Traversing the system stream preferably includes traversing object links of the doubly-linked list of objects. The objects are further preferably cached with an LRU process. Multiple disjoint locations in the system stream can be accessed with a virtual linked list. The virtual linked list can be used to search forwardly, backwardly, or both forwardly and backwardly.

The invention discloses a video image processing method and a related device. The video image processing method comprises the steps of obtaining to-be-processed video image data, and determining a frame supplementing position of the video image data; detecting the position of each preset target in a frame image of the video image data in a bidirectional search mode, wherein the bidirectional searching mode is searching from front to back and from back to front according to the time sequence of the video frames; obtaining a motion track of each preset target based on the position of each presettarget in the frame image; determining the position of each preset target in a to-be-interpolated frame corresponding to the frame supplementing position according to the motion trail to form to-be-interpolated frame image data; and inserting the to-be-inserted frame image data into the frame supplementing position to form the video image data after frame supplementing. Through the scheme, the fluency of the video image formed by the video image data can be improved.

Provided is a system for advertisement management and search through bi-directional search of advertisement objects and advertisement companies. The system includes an information storing unit for storing information on individual or seller members, seller information, goods information, advertisement or event information; a search unit for allowing the advertisement requester to search for the advertisement objects; a selecting unit for allowing the advertisement requester to select the advertisement objects corresponding to an input search condition; a confirming unit for allowing the advertisement requester; and an incentive providing unit for providing incentive to the advertisement objects. The seller activity score as the search result order criterion is effective within a predetermined period which depends on an evaluation criterion item, a business type, and a region, thereby maintaining fairness and reliability of the evaluation result.

The invention discloses an automatic guided vehicle path planning method for routing inspection, which is applied to textile weaving routing inspection. According to the method disclosed in the invention, based on a local bidirectional closed loop and a bidirectional search algorithm, path global optimal planning in textile weaving inspection and path replanning after dynamic obstacle avoidance are effectively solved; the conflict problem of multiple automated guided vehicles in path planning is solved through arrangement and updating of time windows and a conflict negotiation strategy. A proper intelligent system is matched, and therefore the automatic and intelligent transformation and upgrading requirements of a production line in the textile weaving industry can be met.

The invention provides a multi-service call synthetic method in a GSM communication system, which conquers the problem of the prior call synthetic method that: the service type is not distinguished when performing the call synthesize of the GSM network A interface information, thus the difficulty in the monitoring and analyzing process is increased and the work efficiency is reduced. The multi-service call synthetic method in the GSM communication system performs the call synthesize of the multiple concurrence or correlation information alternation of the various service types in a once SCCP connection using a tree-shaped structure. The bidirectional search is performed between father and son CDR and the overall information in the whole connection is seen in the father CDR and the whole information of the father CDR is seen in the son CDR and the position and relation of the each information in the whole information flow (namely father CDR) in the son CDR is seen.

The invention discloses a route planning method based on a road chain, and mainly solves the problems of long search time and many route corners of the existing route planning method. The method adopts the scheme that the method comprises the following steps: extracting node-arc section information of a road network, obtaining relevance between nodes, between the nodes and arc sections as well as between the arc sections, extracting the road chain meeting road consistency by using the node-arc section information, obtaining feasible routes from an origin to a destination by using a bidirectional search route planning method according to road chain information, origin and destination positions input by a navigation system, and selecting the optimal route from origin to the destination by comparing weights of the feasible routes. The method shortens the route search time, reduces the frequency of the route corners, and can be used for traffic control, resource allocation and vehicle navigation, and the obtained route keeps the road consistency.

The invention provides a method and a system for automatic video inserting frame compensation in mediated reality. The method comprises the following steps of adopting a 3D depth camera to collect a real image in display; fusing the real image collected by the 3D depth camera and a virtual scene image through a visual image processing computer; obtaining an engine refresh rate, judging whether the engine refresh rate is lower than a preset frame rate threshold, if yes, storing two frames of fused video data fn-1 and fn, carrying out 8*8 bidirectional search weighting motion estimation and smoothing filtering on the two frames of data, judging whether the two frames of data meets a matching condition, if the matching conditions is met, utilizing a linear interpolation of a 4*4 pixel motion vector to obtain an initial interpolation value, otherwise splitting 8*8 blocks into 4*4 blocks to inherit a result of motion estimation of the last stage and carrying out smoothing filtering. According to the method and the system, the blocking effect caused by block partition in calculation of the motion vector can be effectively improved, and the accuracy of the motion vector is greatly improved.

The invention discloses an urban road network shortest path acquisition method based on angle limitation and bidirectional search. The method comprises the following steps: 1, constructing an urban network according to real-time road condition information; 2, introducing a forward search boundary internal and external intersection set UnF, a backward search boundary internal and external intersection set UnB, a bidirectional boundary intersection set Mn and travel time upper and lower bounds T; 3, updating the intersection sets UnF and UnB inside and outside the forward and backward search boundaries of the sets; 4, obtaining the shortest path from the starting point to the destination point through the intersections in the intersection set Mn of the bidirectional boundaries by a label correction method; and 5, if the travel time of the shortest path is equal to the lower bound T of the travel time or the intersection set UnF and UnB in the forward and backward search boundaries are not updated any more, acquiring the shortest path, or otherwise, returning to the step 3. According to the method, angle limitation and bidirectional search are added in navigation of the urban road network, so that the search range is effectively narrowed; therefore, the navigation efficiency is improved, and a faster and more efficient driving path can be provided.

The invention discloses a robot path planning method fusing a bidirectional search mechanism and an improved A * algorithm. The method aims at solving the problems that a path planned through an A * algorithm has the risk of touching obstacles, and too many redundant points and turning points are generated. The improved A * algorithm fused with the bidirectional search mechanism is provided, firstly, a search direction positioning strategy is introduced, node search is reduced, and algorithm efficiency is improved; secondly, an improved child node expansion strategy is added, so that a safe distance is kept between the robot and the obstacle; a bidirectional search mechanism is fused, and an improved A * algorithm is utilized to search at a target point and a starting point at the same time, so that the planning time is further shortened. An initial path is obtained in a two-dimensional grid map environment, a path turning point is extracted through a vertical distance limit value method, a greedy thought is introduced to select an optimal turning point, the optimal turning point is fitted based on a B spline curve, and a smooth path is obtained. According to the method, the optimal path which is short in time consumption, free of touch and few in turning points can be found.

The invention provides an unmanned aerial vehicle path planning method with endpoint direction and time constraints. The method comprises the following steps: constructing an environment and dividinga space; inputting coordinate positions and corresponding orientation angles of a starting point and a terminal point, and calculating a new starting point and a new terminal point; improving a sparseA * algorithm, increasing acceleration cost in a cost function, adjusting node acceleration cost values with the goal of reaching the end point in the specified time as the criterion, and using the algorithm for path planning; adding the original starting point and terminal point into a path; and carrying out bidirectional search, and selecting an optimal flight path to output for finishing flight path planning. The method is based on the improved sparse A * algorithm, the problem of flight path planning of an unmanned aerial vehicle under complex constraints can be solved, and complex constraints include endpoint direction constraints of departure and arrival at a preset angle, time constraints, unmanned aerial vehicle kinematics constraints and obstacle threat avoidance constraints.

The invention discloses a reliability evaluation method for a distribution network under a maintenance mode based on a deep bidirectional search method, which comprises the steps of S1, collecting andsummarizing data of each typical network structure in a distribution network system to be studied; S2, building models of the typical network structures in the distribution network system based on anMATLAB platform; and S3, carrying out simulated calculation on the model of each typical network structure by adopting the deep bidirectional search method, and obtaining a reliability calculation result of the distribution network system under the maintenance mode through simulation.