72results about How to "Efficient automatic detection" patented technology

The invention discloses a detection method of bituminous mixture, including the following steps: digital image collection of bituminous mixture is completed; image processing of the acquired digital image is completed to obtain the content of compositions of the bituminous mixture. Moreover, the invention also discloses a detection system for the bituminous mixture. Both the method and the system can complete effective separation and extraction of bitumen, airgap, aggregate and aggregate particle of different particle diameter and realize direct determination and calculation of volumetric composition characteristic parameter of the bituminous mixture through converting the obtained characteristic parameter into volume ratio; moreover, the invention improves the mechanical property of the bituminous mixture and provides reliable basis for the scientificity in bituminous pavement engineering research.

The autofocus system of the present invention is characterized by comprising a target subject specifying device which specifies a target subject to be put into focus, a tracking image pickup device which captures a subject image in the shooting range of the camera, and is provided in addition to an image pickup device of the camera which captures a subject image to be recorded or regenerated, a target subject movement detection device which detects a movement in the shooting range of the target subject specified by the target subject specifying device according to an image signal of a subject image captured by the tracking image pickup device, and an AF area change device which changes a range of the AF area such that the movement of the target subject detected by the target subject movement detection device is tracked.

The invention discloses an audio and video conjoint analysis-based fighting detection system and an audio and video conjoint analysis-based fighting detection method thereof. The audio and video conjoint analysis-based fighting detection system comprises a video acquisition module, an audio acquisition module, an algorithm detection module and a management module, wherein each of the video acquisition module and the audio acquisition module adopts a camera, an acquisition card and a computer; the video acquisition module acquires an analogue signal of the camera to a memory of the computer by the acquisition card; the audio acquisition module acquires the analogue signal of the camera to the memory of the computer by the acquisition card; and the algorithm detection module judges the fighting of persons in a target area, judges synchronous audio and video, detects, identifies and positions the chaotic fighting of the persons in the target area, and outputs results to the management module. In the system and the method, the behaviors of the persons in a monitored scene are analyzed, identified and positioned to accurately diagnose the abnormal behaviors such as gathering, fighting and the like of the persons appearing in a video image and give an alarm.

The invention discloses a topic detection or tracking method for network text big data. The basic thought includes the following steps that a graph model of keywords and a corresponding adjacent matrix are built by detecting the keywords occurring in different files, the graph model and the adjacent matrix are combined with spectral clustering to provide a new topic detection model, probability distribution of each file about a topic is calculated, when a new file reaches, similarity between the new file and an attributed set represented by a historical topic is calculated, automatic detection or tracking of the topic is achieved, and a distributed method is achieved through a Map Reduce programming model. The method is characterized in that the topic is displayed and mined through the cooccurrence relation of the keywords rather than an implicit expression, the big data are calculated in a distributed mode, data information in the internet is clustered, expansibility is higher, the quantity of the data capable of being processed is larger, and the throughput rate is greatly increased.

The invention provides a method for identifying faces in videos based on incremental learning of face partitioning visual representations and belongs to the field of pattern recognition. According to the method, an Adaboost algorithm is used for detecting frontal face images in a first frame of the face videos, a Camshift algorithm is used for tracking, all face images are obtained, in the process of reading the face images in the videos, incremental cluttering is carried out on the face images, and a representative image is selected from each kind of face images; the representative images are processed, and a visual dictionary based on the piece visual representations is learnt; the visual dictionary is used for carrying out the representations on the face images; finally, according to similar matrices, the videos composed of the face images are identified. According to the method, an identification rate and robustness of the video faces can be improved under the state that illumination, postures and tracking results are not ideal. The faces in the videos can be detected, tracked and identified effectively, conveniently and automatically.

The invention provides a quick image stain removing method based on the image inpainting technology. Firstly, the camera with a stained lens is used for shooting a white paper image, the Sobel operator and extensive operation are combined to detect the position of the stain, and the white paper image serves as a shade; the shade is used for covering the image shot by the same camera to mark the stain area in the image; for every pixel on the stain area contour, a patch block is defined and the priority of each patch block is calculated; the patch block with the biggest priority is selected to search and find the most matched block to replace and repair; and the repair image information is updated and re-repaired till the whole stain area of the image is repaired. Through the adoption of the image stain removing method, the stained image can be quickly and automatically repaired, and the linear structure of the image can be satisfactorily retained.

The invention discloses full-automatic machine part visual flaw detection equipment. The equipment comprises a feeding device and a detection device, the feeding device is connected with a feeding rail, the detection device comprises a first view finding mechanism, a first conveyor, a turnover mechanism, a second view finding mechanism, a second conveyor, an arrangement mechanism and a third viewfinding mechanism, an output end of the feeding rail faces the first conveyor, the first conveyor is positioned above the second conveyor, the turnover mechanism faces the second conveyor, and the arrangement mechanism is positioned between the second conveyor and the third view finding mechanism. The invention also discloses a full-automatic machine part visual flaw detection method. According tothe equipment and the method, through the cooperation of a vibration disc, the first view finding mechanism, the first conveyor, the turnover mechanism, the second view finding mechanism, the secondconveyor and the third view finding mechanism, comprehensive, efficient and rapid detection on the front surface, the back surface and the side wall of a part is realized, the false detection rate isreduced, the correct and reliable detection result is ensured, the detection efficiency is accelerated, and the cost is effectively reduced.

The invention discloses a weld joint pore defect detection method based on image processing. The method comprises the following steps: carrying out binarization processing on an input pore defect to-be-detected image to obtain a binary image imgb1; performing closing operation and opening operation on the binary image imgb1 in sequence to obtain a binary image imgb2; extracting and putting all connected domains of the binary image imgb2 into a summary set; traversing the summarized set, extracting a connected domain which transversely passes through the image, and putting the connected domain into a first screening set; if only one element exists in the first screening set, taking the connected domain as the contour of the welding joint area, and searching the pore defect area directly; otherwise, searching a steel pipe area; extracting a target welding seam area; searching an air hole defect area; and extracting all pore defect areas based on edge detection, and fitting the contour by adopting a least square method. According to the invention, the to-be-detected image of the pore defect is directly processed, the weld pore defect is detected by combining connected domain searching and edge detection, and the method has good precision and accuracy.

The invention discloses a retinal layering method based on an OCT image. The OCT image of the fundus of the eye is collected as a sample image, the gray-scale normalization processing is carried out on the collected sample images, and then the sample image is flattened. The average filter is used to solve the longitudinal gradient, and then the upper and lower retinal boundaries are determined bycomparison. The calculated features are inputted to the training classifier to obtain the probability that pixels are located at the boundary of different layers. Finally, the optimal boundary and itsposition of the training classifier are optimized by using a graph theory algorithm through the probability. The method of the invention can realize the automatic segmentation of a plurality of layerstructures by using a commonly configured computer, and adopts the multi-level probability classification decision and optimization, and at the same time adopts serialization layering, has fewer original processing steps for the image, and improves detection accuracy certainly.

The invention discloses a cloned subscriber identity module (SIM) card test method, which comprises the following steps that: after authentication requirement is received, whether the address of a network element initiating authentication in the authentication requirement is the same as the registered address of the currently stored SIM card or not is judged, if the two addresses are not the same, a preset analysis rule is matched, and if the matching works, the relative information of the SIM card is obtained and reported to the operator. The invention discloses a cloned SIM card test system at the same time. The method and the system can detect the cloned SIM card in mobile communication network efficiently and automatically and then save labor cost.

The invention discloses a transformer substation equipment detection method and device based on deep learning. The method comprises the following steps: collecting a plurality of images comprising thephysical opening and closing positions of transformer substation equipment, carrying out the labeling and preprocessing of the images, making a standard data set, and dividing the data set into a training set, a verification set and a test set in proportion; and training a substation equipment detection neural network model by using a py master rcnn deep convolutional network, and testing by using the trained network model to obtain an equipment category and a physical opening and closing position recognition result of the substation equipment. According to the transformer substation equipment detection method and device based on deep learning, the transformer substation equipment category and the physical opening and closing state can be efficiently and automatically detected, the recognition result accuracy is high, and the practicability and operability are high.

The invention discloses a hand-held safety-check device and a safety-check method. The hand-held safety-check device comprises a detection main body, the detection main body is provided with a display module and a detection surface, the detection main body also comprises an electromagnetic induction module, a microwave detection module and a master control module, the electromagnetic induction module and the microwave detection module are connected with the master control module, the master control module makes an analytical comparison between an electromagnetic detection signal result output by the electromagnetic induction module and a microwave detection signal result output by the microwave detection module to obtain a detection analysis result, the master control module is then connected with a display unit and performs displaying and early warning on the detection analysis result, and the microwave detection module comprises a microwave scattering imaging algorithm unit.

The invention discloses full-automatic aluminum-plastic bubble cap detection equipment, which comprises a first conveying mechanism; a material blocking mechanism which is arranged at the discharging end of the first conveying mechanism; a first detection mechanism which is arranged above the first conveying mechanism; a second conveying mechanism which is arranged behind the first conveying mechanism; an overturning mechanism which is used for overturning aluminum-plastic bubble caps located at the material blocking mechanism on the first conveying mechanism and then placing the aluminum-plastic bubble caps on the second conveying mechanism; a second detection mechanism which is arranged above the second conveying mechanism; and a sorting mechanism which is arranged at the discharging end of the second conveying mechanism. According to the full-automatic aluminum-plastic bubble cap detection equipment provided by the invention, efficient, rapid and automatic detection of the front and back surfaces of the aluminum-plastic bubble caps and fillers in the bubble caps is realized; the full-automatic aluminum-plastic bubble cap detection equipment greatly reduces the false detection rate of subjective manual visual detection, guarantees the correctness and reliability of a detection result, accelerates the detection efficiency, and effectively reduces the labor cost of detection; the equipment is simple in structure, convenient to install and debug, capable of turning over and adjusting speed and accurate in resetting; the shooting accuracy is high; and the time cost is greatly saved.

The invention discloses a detecting mechanism of an automatic assembling and detecting machine for door locks. The detecting mechanism of the automatic assembling and detecting machine for the door locks comprises a vertical stand column. An upper transverse supporting rod and a lower transverse supporting rod are fixed on the vertical stand column, a vertical mounting plate is fixed on the transverse supporting rod at the upper end, a vertical air cylinder is fixedly arranged at the upper end of the vertical mounting plate, the piston rod of the vertical air cylinder is fixed on a supporting plate, a plurality of vertical supporting columns are fixedly arranged on the lower end face of the supporting plate, the lower ends of the vertical supporting columns are fixed on a square mounting block, a plurality of T-shaped locating rods are inserted in the square mounting block in a penetrating mode, the locating rods are sleeved with compression springs, the upper ends of the compression springs press and butt against the lower end face of the square mounting block, and the upper ends of the locating rods penetrate through the square mounting block and then are fixedly provided with limiting blocks. The detecting mechanism of the automatic assembling and detecting machine for the door locks is capable of effectively carrying out automatic detection and needless for manual detection, reduces detecting errors and allocated operating personnel, and saves labor cost. Meanwhile, the detecting speed is improved through automatic detection, and thus production efficiency is improved.

The invention relates to a surface defect detection method, device and system, and the method comprises the steps: forming spaced stripes on a mirror surface by reflection of a light path, concretely,enabling strip light emitted by a strip light source to be reflected by a first reflector and a second reflector, finally projecting the strip light to a to-be-detected surface of a to-be-detected object, and forming stripes on the to-be-detected surface; and then analyzing by acquiring a stripe image of the to-be-detected surface, and indirectly obtaining a detection result of whether the to-be-detected surface has defects by judging whether stripes formed on the to-be-detected surface are bent or not. By means of the scheme, whether the concave-convex defects exist on the cylindrical surface or the to-be-detected surfaces in other shapes or not can be effectively and automatically detected, the omission ratio is extremely low, manual detection can be replaced, and high detection efficiency is achieved. The scheme has the advantage of high detection reliability when being used for detecting the surface defects of the lithium battery.

The invention provides a method for the size detection and post-treatment of a sheeted molded polymer material, which comprises the following steps of: positioning the sheeted molded polymer material to be detected in a predetermined detection space in a way that the sheet end of one side of the sheeted molded polymer material is positioned at a detection reference point, sensing the position of the other sheet end by using a sensing device to obtain the size of the material, further judging whether the material is oversized or undersized, and performing corresponding post-treatment procedures respectively to increase or decrease the size of the material to a range of allowable error of a good product. In addition, the invention simultaneously provides a detection and post-treatment machine for implementing the method. The method and the machine have the advantages that: automatic detection can be rapidly and highly-efficiently performed on the size of the sheeted molded polymer material, and automatic treatment can be performed according to the size.

The invention discloses a digital screen cabinet modeling design method for production test and a matched test generation method. According to the overall design, components, connector terminals, board cards, devices and logic loops in a simulated real environment in a software virtual environment establish a machine-recognizable digital model for a screen cabinet, a design modeling system establishes a digital model board card through a professional signal digital simulation set of tags, and then a digital device model is established; and finally, establishing a digital screen cabinet model; after modeling, generating a tested screen cabinet label signal set which can be recognized by a machine, and a topological network structure and a logic connection relation between the screen cabinet terminal model and the device terminal model. Therefore, a screen cabinet debugging outline is generated and used for carrying out production testing on the screen cabinet produced by a secondary manufacturer, and the tested screen cabinet which is actually produced after modeling is connected to a detection system through broadband network communication and a modern sensor technology to form a complete closed-loop manual semi-automatic detection system or an intelligent automatic detection system.