11results about How to "Clear division" patented technology

The invention discloses a road crack segmentation method and device based on deep learning, electronic equipment and a program product. According to the method, two collaborative and functional complementary processing paths are adopted to perform parallel processing on a to-be-segmented image so as to obtain corresponding feature information: on one hand, the to-be-segmented image is converted into a feature sequence containing spatial position information, and global semantic modeling is performed on the feature sequence; global semantic features representing the overall shape of the crack and long-range semantic dependence are obtained; and on the other hand, local texture and edge detail information of the crack is directly extracted from the to-be-segmented image, and detail enhancement features are obtained. Thirdly, fusing the global semantic features and the detail enhancement features to obtain target fusion features considering crack positioning, connectivity and boundary and texture expression; according to the segmentation result output based on the target fusion features, even tiny cracks can be effectively depicted, and then the continuity, boundary definition and robustness of tiny crack segmentation under the complex road background can be improved.

The present application relates to the technical field of artificial intelligence education assistance, in particular to an electronic schoolbag assisted learning method and system based on artificial intelligence, which comprises the following steps: collecting voice interaction, touch operation, visual fixation and environmental background signals through an electronic schoolbag sensor array to form a multi-modal original signal stream, separating out a structured learning behavior sequence and a semantic interaction intent through signal decoupling and intent understanding processing, inputting the two types of data into a cognitive state tracking model and a natural language intent analysis engine respectively to generate a real-time cognitive state vector and a structured interaction request, constructing a knowledge subgraph according to the cognitive state vector, combining the interaction request to call adaptive multimedia materials, binding through three-dimensional rendering and interaction logic to generate an immersive learning scene with interactive feedback capability. The method can completely capture the learning state and interaction demand, realize accurate adaptation of learning content and learner state, and strengthen the interactivity and immersion of the learning scene.

The invention discloses a method, a system and equipment for troubleshooting a communication terminal of an all-moving simulator, and belongs to the technical field of aviation simulation equipment fault diagnosis. The method comprises the following steps: acquiring data of a communication terminal operation state, an input / output system state and a central control console environment state, synchronizing time, and constructing a time sequence feature vector through denoising normalization processing; inputting a Kalman filtering state space model to calculate an observation residual error, and distinguishing a fault source and interference characteristics; identifying a fault type by using an integrated decision tree classification model, positioning a target fault point in combination with a fault tree and a Bayesian network, and matching historical cases to generate a troubleshooting scheme; after troubleshooting is completed, the fault case library is updated, and the model is optimized. According to the scheme, intelligent and standardized troubleshooting of faults is achieved, the problems that a traditional method is low in efficiency and poor in accuracy, and fault sources are difficult to distinguish are solved, the troubleshooting efficiency and accuracy are greatly improved, the operation and maintenance cost is reduced, and stable operation of the full-motion simulator is guaranteed.

The utility model discloses a kind of brining devices, belong to food brining technical field, including hanging conveying mechanism and several brining pots, hanging conveying mechanism includes frame body, chain, and frame body is connected and installed with guide rail, driving mechanism;Chain is hung with lifting device;Still include several heating mechanisms, pouring mechanism.When using, raw material is moved to front side by being added in rear brining pot, and brining pot is heated by falling on heating mechanism after brining pot descending;After brining, brining pot is lifted, and moves to rear side, and brining pot falls, and the pouring mechanism is realized to pour brining pot;And then the next cycle of brining operation can be realized.Multiple brining pots can be realized continuous heating brining, automatic pouring operation, can greatly improve product brining efficiency, reduce manual intervention, reduce labor intensity, greatly improve single brining amount, ensure the stability and consistency of product quality, can be fully adapted to large-scale brining production.

The application provides a kind of multilingual front-end resource automatic construction method and system based on construction stage, method includes: in front-end construction stage, configuration target language type set and main language type, and target language type set is executed cyclic construction processing to;In cyclic construction processing, when target language type is different from main language type, the text content in front-end source code is scanned and identified, generates and labels text mark, based on text mark, from target language type corresponding language package, target language text is obtained, and target language text is replaced with corresponding text content, and the text content that is not completed replacement processing is recorded as to-be-processed text;Call translation interface to execute translation processing to to-be-processed text, generate corresponding translation text, based on translation text, corresponding text content is executed replacement processing, and according to the text mark of to-be-processed text, translation text is added to target language type corresponding language package.

The invention discloses a long-tail sonar image open set identification method and system based on class equalization OpenMax, and relates to the technical field of sonar image identification. The method comprises the following steps: carrying out five-fold cross division on a data set for five times according to a preset known class number and an unknown class number; training a feature indicator to extract a known class sample activation vector, and calculating a mean activation vector and the distance between the sample and the center; determining the number of tail samples of each category in a self-adaptive manner through a category reward or intra-category distance distribution strategy, and fitting a Weibull model; during testing, the distance between a sample and the center of the known class is calculated, the softmax probability is corrected in combination with the survival probability of the Weibull model, and recognition of the known class and the unknown class is achieved. The method solves the problem of modeling instability caused by traditional OpenMax fixed tail size, gives consideration to known class recognition precision and unknown class detection capability, is adaptive to long-tail sonar data distribution, and is suitable for underwater open scenes.

This application discloses a digital calibration circuit and calibration method for a temperature-sensitive sensor, including a temperature acquisition module, a non-volatile storage module, a computation module, an output processing module, and a finite state machine for coordinating the time-sharing operation of each module. The temperature acquisition module is configured to acquire the analog temperature signal output by the temperature sensor and convert it into an N-bit temperature raw code. The non-volatile storage module stores multiple sets of temperature boundary codes and their corresponding calibration values ​​in ascending order of temperature. The computation module is configured to, under the scheduling of the finite state machine, first coarsely locate the target interval through sequential traversal, and then converge to obtain the final calibration value through synchronous binary search iteration. The output processing module is configured to process the final calibration value and output a calibration signal to the sensor's analog front end. This application adopts a processorless pure hardware architecture, does not rely on a CPU or software, has fewer circuit gates, and significantly reduces chip area and power consumption.

A kind of mineral control filter element includes: filter element shell;Center tube;Water distributor;Mineral containing basket;The utility model discloses a filter element structure is provided, which can realize clear mineral partition, uniform water flow regulation and control, and controllable mineralization ratio, and can realize stratified flow of water and accurate utilization of mineral by specific water distribution assembly and structure of mineral partition placement, thereby stabilizing the adjustment of water purification mineralization ratio and concentration, improving mineralization efficiency and taste consistency.

This invention discloses a method, apparatus, computer equipment, and storage medium for predicting the organic matter content of source rocks, relating to the field of oil and gas exploration. The method includes: acquiring seismic and well logging data; interpreting the seismic data based on sequence interpretation technology; processing and calculating well logging curve data; predicting the TOC data of the target layer using a seismic inversion algorithm; and obtaining the distribution of source rocks in the target layer based on the TOC inversion data. This invention also provides a method, apparatus, computer equipment, and storage medium for predicting the organic matter content of source rocks. This invention fully considers the core measurement data, well logging data, and seismic response characteristics of source rocks, links well logging characteristics and seismic characteristics through well-seismic calibration, and predicts TOC through a joint well-seismic inversion method, thereby predicting the distribution of high-quality source rocks.

The utility model discloses a double-cooling modular soft starter. By optimizing the internal structure of the base and setting a partition tabletop to partition the conductive busbars and the radiator, on the one hand, the use safety between the current-carrying components and the main heat-dissipating components inside the soft starter is ensured. The fixing plate of the partition tabletop is used for the fixed connection of the radiator. A heat dissipation grid is specially opened on the connecting plate for the current transformer, and the existing heat dissipation fan assembly of the soft starter is used to directly and quickly discharge the internal heat. On the other hand, the working temperature of the current transformer is significantly reduced. By effectively reducing the working environment temperature of the current transformer, it can work stably within the allowable temperature range, thereby greatly improving the accuracy and stability of current sampling and significantly enhancing the operation reliability and control accuracy of the entire soft starting system.

The utility model discloses a multi-surface reflection three-dimensional signal lamp structure, and belongs to the technical field of vehicle lamp structures. Comprising a thick-wall frame, a light source set and a thick-wall piece, the thick-wall frame is a polygonal frame, and the inner wall of the thick-wall frame is subjected to an aluminum plating process; the light source set is fixedly installed on the rear surface of the thick-wall support, the light source set comprises a PCB, the front surface of the PCB is fixedly connected with the thick-wall support, and an LED lamp is arranged on the front surface of the PCB; the thick-wall part is fixedly installed in the thick-wall frame, the light-in face of the thick-wall part is located in front of the LED lamp, and the light-out face of the thick-wall part is located in the thick-wall frame. According to the multi-surface reflection three-dimensional signal lamp structure, the aluminized thick-wall support is arranged, so that the luminous thick-wall pieces respectively present a virtual image on the aluminized inner wall, the diamond-feeling three-dimensional effect is formed, the requirement of consumers for the appearance of a vehicle is met, and meanwhile the requirement for uniform light of a vehicle lamp is met.