39results about How to "Capture comprehensive" patented technology

The invention discloses a flight track prediction method, which considers the meteorological environment where the flight is located, uses a convolutional neural network to extract meteorological features, uses a cyclic neural network to model the track features, and finally outputs a series of future flights Trajectory, specifically including track data preprocessing, meteorological feature extraction, definition of time series model, model compilation and training, track prediction. Through the technical scheme of the invention, the long-term trajectory of the flight can be accurately predicted in the actual environment.

The invention discloses a structure provided with multiple rings of annular shaver net bodies and capable of forwards rotating and reversely rotating. The structure comprises a shaver net which is commonly constituted by the first shaver net body 1, the second shaver net body 2, the third shaver net body 3 and the multiple rings of shaver net bodies, the second shaver net body 2 is arranged in thefirst shaver net body 1, and the third shaver net body 3 is arranged in the second shaver net body 2; the first shaver net body 1, the second shaver net body 2, the third shaver net body 3 and the multiple rings of shaver net bodies forwards rotate or reversely rotate with a mounting shaft 5 as a symmetry point, and inner blades are arranged in the first shaver net body 1, the second shaver net body 2, the third shaver net body 3 and the multiple rings of shaver net bodies correspondingly; and a plurality of beard inlet grooves 4 are formed in the upper surfaces of the first shaver net body 1, the second shaver net body 2, the third shaver net body 3 and the multiple rings of shaver net bodies correspondingly. The arranged multiple rings of annular shaver net bodies and beard inlet grooves can forwards rotate and can also reversely rotate, thus the shaver net can actively capture beard, the shaving efficiency is higher, the structure is simple, and good using prospects are achieved.

The invention discloses a photographing device for industrial visual inspection, which comprises a bottom plate, two fixed sleeves are arranged on the upper surface of the bottom plate through a moving frame, and inside the two fixed sleeves are provided with a device for changing the placement position of workpieces. Two sets of support and deflection mechanisms; each set of support and deflection mechanisms includes a rotating sleeve and a fixed plate, the rotating sleeve is rotatably mounted on the inner wall of the fixed sleeve, the fixed plate is fixedly mounted on the upper surface of the bottom plate, and the upper part of the fixed plate A second gear is rotatably installed, and the support plate is arranged inside two rotatable sleeves, one of which is equipped with a camera for taking pictures of the workpiece in different directions. Through the setting of the support deflection mechanism, the invention can realize taking pictures of different positions of the workpiece, so that the image of the object can be captured more comprehensively and the detection quality can be improved.

The facial expression capture method, system and equipment disclosed in the present invention relate to the field of computer vision technology, and specifically involve the use of face detection technology, key point detection technology and facial motion fitting technology to obtain facial motion information in real time, and then display it through a deformation model . The facial expression capture method, system and equipment proposed by the present invention utilize rich deformation models and advanced facial motion fitting technology to capture every subtle change of the face as much as possible and improve the computing efficiency to the greatest extent. .

The invention discloses a multi-parameter analysis sample holder comprising a water inlet pipeline, a circulation bucket, a drainage pipeline, a sensor, a water outlet valve, an air pump, a sampling pipeline, a sample holding valve, a sampling pump, a sample mixing bucket and a sample holding bottle, thereby analyzing and detecting sewage in a drainage pipe. An inlet valve and a circulation pump are arranged on the water inlet pipeline; the circulation bucket and the water inlet pipeline are communicated via a sampling valve; the drainage pipeline and the circulation bucket are communicated via a pressure release valve; the air pump is connected with the circulation bucket via an exhaust moving valve; the sampling pipe is connected with the water inlet pipeline via the sampling valve; thesample mixing bucket is connected to the tail end of the sampling pipeline; and the sample holding valve and the sampling pump are arranged at the sampling pipeline. The sample holding bottle is communicated with the sample mixing bucket by the sample holding valve. Therefore, the sampling flexibility is improved; the supervision cost is lowered effectively; the supervision level is improved; andthe environmental protection is promoted.

The invention discloses an ICU death risk assessment system based on deep learning, which includes: ICU historical database, which stores the historical patient's sign data set and the historical patient's real final state; the first data preprocessing module, which is obtained from the Extract the historical patient's sign data set from the ICU historical database and perform preprocessing to obtain training sample data, and simultaneously extract the real final state of the patient and assign a label to the training sample data; the second data preprocessing module, which is extracted by The human-computer interaction module enters and preprocesses the sign data of the patient to be evaluated; the death risk assessment module is constructed based on a two-way supervisory LSTM neural network; it acquires training sample data and labels from the first data preprocessing module Values ​​are used for model training, and the trained model is used to obtain the sign data of the patient to be evaluated from the second data preprocessing module for evaluation, and finally the evaluation result is output through the human-computer interaction module.

The invention belongs to the field of user modeling and recommendation systems, and specifically relates to a meta-path-based context-aware user modeling method and a sequence recommendation method, aiming to solve the problem that the existing user modeling methods only consider the simple sequence relationship of user historical behavior, Complex user preferences cannot be captured; the accuracy of the existing recommendation system recommendations is not high, and the recommended content does not have diversity. The present invention acquires local interest representations based on the sequence of items accessed by users, obtains context-aware global interest representations by modeling heterogeneous networks, and combines local interest representations and global interest representations to obtain user comprehensive interest representations. Based on the user's comprehensive interest representation and the items to be predicted, the recommendation results are obtained through the recommendation model. The user modeling method of the present invention can more comprehensively capture user interests, and is beneficial to improving the accuracy of the recommendation system; the meta-path-based context-aware sequence recommendation method has high recommendation accuracy and more diverse recommendation contents.

The invention belongs to the technical field of oil product detection, and discloses an edible oil product color and luster, transparency and impurity detecting device. The device comprises an oil product conveying unit, a backlight bright field detecting unit, a top light dark field detecting unit, a polarized light dark field detecting unit, a detection control and image processing unit and a color and luster and transparency detecting unit, wherein the backlight bright field detecting unit obtains backlight bright field detection image information, the top light dark field detecting unit obtains top light dark field detection image information, the polarized light dark field detecting unit obtains polarized light dark field detection image information, the detection control and image processing unit is connected with the oil product conveying unit and the detecting units, controls actions thereof and processes the obtained detection image information, and the color and luster and transparency detecting unit projects detection signal light from one side of an oil product to be detected through a backlight source, collects light emitted from the oil product to be detected, and analyzes the spectral intensity and feature. The oil product to be detected is conveyed to the detection units in sequence through the oil product conveying unit, and the polarization direction of polarization detection light is perpendicular to that of a polarization analyzer. The edible oil product color and luster, transparency and impurity detecting device is efficient and high in reliability.