40 results about "Motion History Images" patented technology

The invention discloses a motion history image and convolutional neural network-based behavior identification method. The method comprises the following steps of S1, obtaining an input original videoimage, and processing the input original video image through a motion history image-based behavior sequence feature extraction method; and S2, performing behavior identification on a local motion history image by adopting a deep convolutional neural network-based method to obtain a behavior type classifier, and finally outputting a behavior identification result through the behavior type classifier. The motion history image is calculated in an original video sequence, so that the to-be-processed information amount is reduced and key time-space information in behavior identification is extracted; and by taking the motion history image as an input, a deep convolutional neural network is established, then the network is trained by utilizing a stochastic gradient descent method and a Dropout policy, and finally behavior type identification is realized. The method can be effectively applied to online real-time behavior identification.

The invention provides an accurate and rapid road violation and parking full-automatic snapshot system. The system comprises a front-end monitoring point part, a network transmission part and a centermanagement unit. The vehicle detection technologies mainly comprise the vehicle dynamic detection technology and the vehicle tracking technology. A matched vehicle tracking method is estimated by utilizing the motion detection method based on time-based motion historical images and feature-point optical flows. In order to achieve a better motion detection effect, a shadow removing method is usedand a moving vehicle is more accurately detected. The multi-feature point optical flow detection tracking mode is adopted, and the tracking accuracy is obviously improved in combination with a vehicleshielding solution. The system is powerful in function and is sufficient to meet the requirements of various application occasions. By adopting the system, the functions of the remote control on a ball machine at a front end monitoring point, the previewing of the video monitoring, the automatic detection of the illegal parking, the automatic identification of license plate numbers, the association of the high-definition video recording with illegal behavior videos, the storage and the browsing of the vehicle information, the system management, the remote maintenance and the like are realized.

The invention discloses a view-independent human action identification method based on template matching, which can identify a plurality of pre-defined typical actions in a video. When constructing a template, a motion history image under a plurality of projection viewpoints are calculated for each sample action and polar coordinate characteristics are extracted, the polar coordinate characteristics are mapped to a low-dimensional sub-space by adopting a manifold learning method, and super balls are constituted for the sample actions in the sub-space on the basis of the low-dimensional coordinate of the multi-viewpoint polar coordinate characteristics. An action template is composed of a plurality of super balls with known ball centers and radiuses. When an unidentified action is given, the motion history image and the corresponding polar coordinate characteristics of the action are firstly calculated, then the polar coordinate characteristics are projected into the template action sub-space to obtain the low-dimensional coordinate, the distances from the coordinate to all the ball surfaces of the super balls are calculated, and the nearest super ball is selected as the identification result. The technology provided by the invention realizes the view-independent action identification and has higher application value in the video monitoring field.

The invention provides a three-dimensional gesture action recognition method based on depth images. The three-dimensional gesture action recognition method comprises the steps of acquiring the depth images including gesture actions; dividing a human body region corresponding to the gesture actions from the images through tracking and positioning based on quick template tracking and oblique plane matching to obtain a depth image sequence after the background is removed; extracting useful frames of the gesture actions according to the depth images after the background is removed; calculating three-view drawing movement historical images of the gesture actions in the front-view, top-view and side-view projection directions according to the extracted useful frames; extracting direction gradient histogram features corresponding to the three-view drawing movement historical images; calculating relevance of combination features of the obtained gesture actions and gesture action templates stored in a pre-defined gesture action library; using a template with largest relevance as a recognition result of a current gesture action. Therefore, three-dimensional gesture action recognition can be achieved by adopting the three-dimensional gesture action recognition method, and the three-dimensional gesture action recognition method can be applied to recognition of the movement process of simple objects.

The invention discloses a behavior recognition method based on depth and RGB information and multi-scale and multidirectional rank and level characteristics. The behavior recognition method comprises steps of: video preprocessing, target motion changing process description, multi-scale and multidirectional rank and level characteristic extracting, model constructing, and model selection and deduction. By means of the behavior recognition method, depth images are used for performing behavior recognition so as to overcome difficulties occurred in visible light image behavior recognition, like interference of lighting changing, shadows, object shielding and the like; secondarily, by means of the method, depth difference value motion historical images and depth limitation RGB image difference value historical images can well capture human body behavior changing processes in image sequences and RGB image sequences, thirdly, the multi-scale and multidirectional rank and level characteristics disclosed by the method have space resolution capability and detail description capability and have good robustness and distinguishing performance; finally, models can be selected independently according to degree of light, and adaptability of a behavior recognition algorithm can be further improved.

The invention provides an image-type fire detection method based on a cumulative prospect image, builds the cumulative prospect image, and is used for real-time detection of a fire image. The method comprises the following steps of: extracting the cumulative prospect image obtained by a monitoring camera through a computer; blocking the image; counting brightness values of all pixels of each image block in the cumulative prospect image; judging according to preset sensitivity; and sending out a warning signal if a fire is detected, otherwise returning to the initial steps to continuously process next frame image. According to the method, the definition of historical motion images is changed to obtain the cumulative prospect image, and the cumulative prospect image can be used for reflecting active characteristics of fire flame better and greatly reducing false alarm rate and has good anti-noise capability. Moreover, the method has the advantages of simpler arithmetic principle, low computation and quite good real-time performance and can meet the requirement of the conventional image-type fire detection technology for the real-time performance.

The invention relates to a motion-detection-based human body abnormal behavior detection method for scene detection of a fixed camera. Motion detection is carried out by using a way of combination of a background subtraction method and a motion historical image, a motion foreground is extracted and background updating is carried out by sing a mask obtained by using a frame difference method, and post processing operation is carried out and a foreground of a human body is filtered; with a method based on context information matching between video frames, and human bodies in a previous frame and a current frame are matched; with combination of an originally-set abnormal condition, system decision is carried out by using geometrical features of a rectangular frame and a center of mass, whether the human body in motion has an abnormal behavior is determined, and an invasion direction is determined; and then counting of the abnormal people and human body track marking after abnormal behaviors are carried out, thereby achieving an objective of a video analysis. According to the invention, the algorithm complexity is low and the portability is high; and the real-time performance is good and the detection rate accuracy is improved.

A method for object analysis using motion history is provided. The method includes receiving video data comprising a plurality of frames of a scene comprising one or more elements, and processing the video data to produce a motion history image comprising motion history values of at least one of the elements. The method also includes identifying the one of the elements for further processing if a characteristic of the motion history image satisfies a criteria.

The invention provides a motion characteristics extraction method and device. The motion recognition accuracy and robustness can be improved. The method comprises the following steps: obtaining three-dimensional human skeleton data; carrying out storage on tissues of a skeleton model through a tree structure under a local coordinate system according to the obtained three-dimensional human skeleton data, and building a limb tree model; and combining a motion history image with a motion energy image according to the built limb tree model to obtain an Hu invariant moment of describing human motion characteristics. The device comprises an obtaining module, a building module and a motion characteristics extraction module, wherein the obtaining module is used for obtaining the three-dimensional human skeleton data; the building module is used for carrying out storage on the tissues of a skeleton model through the tree structure under the local coordinate system according to the obtained three-dimensional human skeleton data, and building the limb tree model; and the motion characteristics extraction module is used for combining the motion history image with the motion energy image according to the built limb tree model to obtain the Hu invariant moment of describing the human motion characteristics. The motion characteristics extraction method and device are suitable for the technical field of mode recognition.

Methods of processing video are presented to generate signatures for motion segmented regions over two or more frames. Two frames are differenced using an adaptive threshold to generate a two-frame difference image. The adaptive threshold is based on a motion histogram analysis which may vary according to motion history data. Also, a count of pixels is determined in image regions of the motion adapted two-frame difference image which identifies when the count is not within a threshold range to modify the motion adaptive threshold. A motion history image is created from the two-frame difference image. The motion history image is segmented to generate one or more motion segmented regions and a descriptor and a signature are generated for a selected motion segmented region.

Described is a system for detecting moving objects using multi-frame motion history images. An input video sequence of consecutive registered image frames is received. The sequence of consecutive registered image frames comprises forward and backward registered image frames registered to a coordinate system of a reference image frame. Frame differences are computed between each of the consecutive registered image frames and the reference image frame. The frame differences are accumulated based on characteristics of the input video sequence to compute a corresponding motion response value. A selected threshold value is then applied to the motion response value to produce at least one binary image used for detection of moving objects in the input video sequence. Additionally, the invention includes a system for adaptive parameter optimization by input image characterization, wherein parameters that are based on characteristics of the image influence the motion detection process.

The invention provides a hand waving detection method based on motion historical images. The area of hand waving detection is indirectly confirmed through pedestrian detection firstly, then the principal motion direction of the area is calculated through the motion historical images, and finally the motion of hand waving in the sequential images is judged. The method is low in complexity, high in detection accuracy and suitable for the target confirmation phase before target following photographing of the unmanned aerial vehicle.

The invention discloses a specific action detection method for intelligent experiment evaluation. The specific action detection method comprises the following steps: acquiring an operation video of a related experiment by using a camera; fusing continuous n frames of images of the operation video through a motion historical image method, and expressing a target motion condition in an image brightness form to obtain a fused picture; carrying out target detection labeling on the fused picture, and dividing a data set into a training set, a verification set and a test set; the data set is sent to a pre-training model for training, and a target detection model is obtained; and performing action detection on the fused picture of the related experimental operation video by using the target detection model, and evaluating a related score. According to the method, the robustness of intelligent experiment evaluation can be greatly improved, in addition, a video classification task is converted into a target detection task, and the labeling cost is greatly reduced.

An image processing apparatus includes: a camera configured to generate a captured image by capturing motion of an object; an image processor configured to process the captured image; a storage configured to store a predetermined information which is common data included in common in a plurality of motion history images (MHI) obtained by capturing respectively the motions by a predetermined form; and a controller configured to determine that a form of the motion by the object within the captured image corresponds to the predetermined form if it is determined that MHI data of the captured image includes the predetermined information.

A gesture judgment method used in an electronic device having frame capturing function is provided. A plurality of MHI (motion history image) angles/directions are obtained from a plurality of corresponding MHIs. Whether a current gesture control is valid is judged according to the MHI angles/directions. If the current gesture control is valid, then weight assignment is performed on the MHI angles/directions to obtain a judgment result of the current gesture control.

The invention discloses a human body behavior identification method and system based on a motion history image, and the method comprises the steps of extracting a video image frame based on a human body motion video image, and dividing the attribute of each pixel point into a foreground pixel point and a background pixel point through an inter-frame difference method; obtaining a pixel point attribute change sequence at the same position; improving the motion history map MHI to obtain a video image grey-scale map under each change mode; performing feature extraction by taking the grey-scale map under each change mode as a feature image; inputting into a preset classifier model for action behavior identification. According to the invention, an improved motion history map MHI is adopted, theMHI is extracted by using the MHI, the spatial domain information in the MHI is extracted by using 2DHaar wavelet transform, and the time domain information in the MHI is extracted by using the statistical histogram of the MHI, so that the operation complexity is reduced, the feature contains richer motion information, and an identification algorithm with a simple classification process and a high speed is realized.