10results about How to "Reduced post-processing time" patented technology

The invention discloses a multi-angle video automatic editing method based on content defect detection, and relates to the technical field of video processing. According to the method, multi-track video materials are accurately synchronized to the same time axis through cross-correlation calculation, and flaws in video content are analyzed frame by frame and classified and identified by applying fuzzy detection of a dynamic threshold and combining jitter detection and shielding detection of rotation and displacement estimation. And according to a detection result, an editing scheme is intelligently generated through a dynamic programming algorithm of state transition optimization, high-quality fragments are selected, defective parts are automatically removed, the continuity and multi-view diversity of video contents are reserved, and finally, a high-quality recorded video is generated through integration. The video editing efficiency is greatly improved, the manual intervention requirement is reduced, meanwhile, it is ensured that the quality of the output video meets the professional application requirement, and the method is suitable for multiple scenes such as conference recording, educational training and multi-angle video production and has high practical value and popularization prospects.

The utility model discloses a loose fiber acid dyeing device, and relates to the technical field of loose fiber dyeing. The device comprises a base, a supporting frame is fixedly connected to the top of the base, a controller is fixedly connected to one side of the supporting frame, a dyeing barrel is fixedly connected to the top of the base, a supporting plate is installed in the dyeing barrel, a yarn cage is installed at the top of the supporting plate, and a central liquid drainage pipe is fixedly connected to the interior of the dyeing barrel. According to the yarn cage, the dewatering assembly is arranged, after loose fiber dyeing is finished, a valve can be opened to discharge internal dyeing liquid from a liquid outlet, then the dewatering assembly is started to descend, the loose fibers in the yarn cage are extruded through the dewatering assembly, water in the loose fibers is extruded out, and preliminary dewatering is carried out; and meanwhile, subsequent treatment is facilitated, the subsequent treatment time is shortened, and the practicability of the device is greatly improved.

The invention relates to an end-to-end traffic light multi-task sensing method, and the method comprises the steps: enabling a plurality of different tasks to share a model backbone network, and enabling the model parameters to be highly reused; performing weighted summation on the loss of each task to obtain final multi-task loss, and realizing mutual supervision among multiple tasks; in the training stage, a topk10 and topk1 double-branch positive sample distribution strategy is used, and topk10 provides more positive samples for the model, so that training is easy to converge; only a topk1 branch export model is used, so that a model output result is reduced to a lower order of magnitude, complexity caused by using an NMS when multiple targets are used is avoided, and traffic light perception post-processing is accelerated. Three different tasks of traffic light target detection, attribute identification and countdown key point positioning in traffic light perception can be unified into one network model. Traffic image input and single-model reasoning output of a multi-task result are realized. And compared with a multi-model serial mode, the model parameter quantity and the calculation quantity are greatly reduced, and the method is more friendly to embedded equipment.

The invention discloses a wear-resistant aluminum alloy and a preparation method thereof, and relates to the technical field of aluminum alloys. The intermediate alloy and the aluminum ingot are added at the same time, so that the melting and dissolving time can be prolonged, and high-melting-point elements can be dispersed in molten aluminum; by means of mechanical alloying and rapid sintering, the surface of the aluminum alloy is covered with wear-resistant particles, the wear resistance of the aluminum alloy is improved, meanwhile, the mechanical performance of the aluminum alloy is guaranteed, and metallurgical bonding is formed after subsequent sintering; mechanical alloying can improve the mixing uniformity between the aluminum alloy powder and the wear-resistant particles and the bonding energy between the wear-resistant particles and the aluminum powder; after the wear-resistant layer is prepared, through heat treatment, the form in the alloy can be remarkably refined, and a strengthening phase is dispersed in a matrix; according to the aluminum alloy material produced through the technology, high-quality aluminum alloy can be forged, high strength is achieved, good abrasion resistance is achieved, and the post-treatment time of the aluminum alloy can be saved.

The invention relates to a flat knitting flaky stab-resistant structure fabric and a preparation method thereof, and relates to the field of textile materials. The fabric is of a hollow stab-resistant structure which is regularly arranged and overlapped, and a minimum cycle organization unit of the hollow stab-resistant structure comprises a surface layer, a bottom layer and a middle layer located between the surface layer and the bottom layer; a plurality of bag-shaped hollow interlayers with three sides connected and one side opened are formed between the surface layer and the middle layer and between the bottom layer and the middle layer through needle turning actions, and puncture-proof blocks are arranged in the bag-shaped hollow interlayers in a replaceable mode; wherein the surface layer is provided with an extending part, and the extending part is located in the direction close to the opening side of the bag-shaped hollow interlayer. The fabric can be integrally knitted and formed only by a four-needle-bed computerized flat knitting machine, the area of a surface layer structure unit can be larger than that of a middle layer (bottom layer) structure unit by adopting different knitted structures and changing the wale density of the surface layer, the middle layer and the bottom layer of the fabric, so that the stab-resistant blocks can be better attached to the interior of the knitted fabric, and falling of the stab-resistant blocks is avoided.

The invention relates to a method for detecting objects and key points. The method comprises the following steps of: receiving an image; executing a deep neural network architecture based on the image to obtain one or more object bounding boxes; executing a deep neural network architecture based on the one or more object bounding boxes to obtain relative positions of one or more key points corresponding to the one or more object bounding boxes; and outputting the positions of the one or more object bounding boxes and the one or more key points.

The utility model relates to the field of annealing devices, in particular to an annealing device for a photovoltaic solder strip double-end tinning machine, which is a double-end annealing device and comprises at least even sets of annealing mechanisms, one set of annealing mechanism is a first annealing mechanism, the other set of annealing mechanism is a second annealing mechanism, and the first annealing mechanism and the second annealing mechanism are arranged side by side. The first annealing mechanism and the second annealing mechanism operate independently or simultaneously; the first annealing mechanism comprises a first pay-off wheel, and a first wire sequentially passes through the first annealing wheel, a first annealing water tank, a first nitrogen glass pipeline and a first air blowing assembly from the pay-off position of the first pay-off wheel and then enters the next procedure through a second baffle; a second wire sequentially passes through the first baffle, the second annealing wheel, the second annealing water tank, the second nitrogen glass pipeline and the second air blowing assembly from the pay-off position of the second pay-off wheel and then enters the next procedure, the first baffle is arranged on the upper portion of the first annealing mechanism, and the second baffle is arranged on the upper portion of the second annealing mechanism. The utility model has the advantages of simple structure and reasonable layout, can effectively shorten the follow-up processing time after the wire is broken, and improves the production efficiency.

The invention provides a method for efficiently preparing a furalan key intermediate, which comprises the following steps: (a) adding 2-methyl-4-acylbenzoic acid, 3, 5-dichlorotrifluoroethyl acetophenone, an organic solvent and triethylamine into a reaction container, heating to 50-60 DEG C, and reacting for 8-24 hours to obtain a first reaction mixed solution; (b) cooling the first reaction mixed solution to 30-40 DEG C, adding 4-dimethylaminopyridine, then dropwise adding acetic anhydride, and carrying out heat preservation reaction for 3-5 hours to obtain a second reaction mixed solution; and (c) cooling the second reaction mixed solution to-5-0 DEG C, adding alkali and hydroxylamine hydrochloride, carrying out a heat preservation reaction for 1-2 h, recovering the organic solvent at 50-60 DEG C under negative pressure, adjusting the acid, crystallizing with ethanol, and filtering to obtain the Frerala key intermediate 9. The target product is synthesized by a three-step reaction one-pot method (without intermediate treatment), the post-treatment is simple and efficient, the product yield is stable, and the purity content is high.

The invention belongs to the technical field of flight parameter data processing, and particularly provides a method for judging the steady state of flight parameter data, which comprises the following steps: extracting flight parameter point location data to generate a parameter list; calculating and combining time intervals meeting preconditions; screening parameter points of the main parameters in the merging time interval; two stable time intervals are obtained through comparison of adjacent point thresholds and comparison of interval time thresholds; and taking the intersection of the two intervals to obtain a final steady-state time interval. According to the method, the sample size is reduced through precondition screening, the data processing efficiency is improved, irrelevant data is filtered, the judgment credibility is improved, and the method can be used for deducing pilot operation habits, judging the sensitivity of finished products / structural parts and predicting system faults and is suitable for flight parameter data steady-state analysis of various airplanes.

This invention discloses an automatic vehicle modal extraction method and related equipment. The method includes: obtaining basic order information of the first-order torsional mode and the first-order bending mode corresponding to the basic model, and basic model displacement information of the target point, respectively, through modal simulation analysis; obtaining modal calculation information corresponding to multiple parametric models through simulation analysis; calculating the similarity information between the displacement information of the parametric models corresponding to different parametric models and the displacement information of the basic model; and obtaining the torsional frequency information and torsional order information of the parametric models corresponding to the first-order torsional mode, and the bending frequency information and bending order information of the parametric models corresponding to the first-order bending mode, based on the similarity information. The method provided in this application can significantly reduce the post-processing time of simulation results, achieve automatic identification of modal information of parametric models, and improve simulation efficiency in the design phase.