146 results about "Object-class detection" patented technology

The invention discloses an object detection and tracking method of fusing laser point clouds and images. The method comprises the steps of: S1, respectively collecting laser point cloud data and imagedata of an object, carrying out first object detection according to the collected point cloud data to obtain a first object detection result, and carrying out second object detection according to thecollected image data to obtain a second object detection result; S2, carrying out fusion judgment on the first object detection result and the second object detection result on the basis of Bayesiandecision to obtain final object detection result output; and S3, carrying out object tracking according to a final object detection result. The method has the advantages of a simple and highly-efficient realization method, high object detection and tracking accuracy, high environment adaptability and fault tolerance, stability, reliability and the like.

Techniques are disclosed for detecting foreground objects in a scene captured by a surveillance system and tracking the detected foreground objects from frame to frame in real time. A motion flow field is used to validate foreground objects(s) that are extracted from the background model of a scene. Spurious foreground objects are filtered before the detected foreground objects are provided to the tracking stage. The motion flow field is also used by the tracking stage to improve the performance of the tracking as needed for real time surveillance applications.

The present invention relates to an object detection and behavior recognition system using three-dimensional motion data. The system receives three-dimensional (3D) motion data of a scene from at least one sensor, such as a LIDAR sensor. An object is identified in the 3D motion data. Thereafter, an object track is extracted, the object track being indicative of object motion in the scene over time. Through Dynamic Time Warping (DTW) or other comparison techniques, the object track is compared to a database to identify the behavior of the object based on its object track.

The invention discloses a small target semantic segmentation method combined with target detection. Attention semantic segmentation network, training the network to get the whole semantic segmentationmodel; making small target detection dataset and small target semantic segmentation dataset; training the small target detection network based on YOLOv2 through the small target detection data set; asmall target semantic segmentation network is designed and trained by using the small target semantic segmentation data set to obtain the small target semantic segmentation model. In the testing phase, the test image is used as the input of the whole semantic segmentation model and the small target detection network, and the segmentation result and the small target boundary box of the whole imageare obtained, which is modified by the small target semantic segmentation model. The invention can greatly reduce the segmentation difficulty of the small target, thereby effectively improving the segmentation performance of the small target.

The invention discloses a surveillance video abstraction extraction method based on moving object detection. The surveillance video abstraction extraction method comprises the steps of firstly determining the parameters of a ViBe model, initializing the model by use of a section of video frames and treating new video frames by use of the obtained ViBe background model so as to generate binary background images, secondly, performing morphological operation on each obtained binary image and obtaining the position information of a moving object in the frame of the binary image, thirdly, tracking each moving object and storing the motion sequence of each object, fourthly segmenting the moving objects to obtain fine profiles of the objects, and finally, jointing the profile sequences of the motion objects to the background frames by use of a reasonable method. A simple and effective video abstraction extraction method is realized by use of the ViBe model for various surveillance videos, and the processing efficiency is improved on the basis of remaining the motion information of the key moving objects.

A multi-stage method of visual object detection is disclosed. The method was originally designed to detect humans in specific poses, but is applicable to generic detection of any object. A first stage comprises acts of searching for members of a predetermined general-class of objects (such as humans) in an image using a cognitive swarm, detecting members of the general-class of objects in the image, and selecting regions of the image containing detected members of the general-class of objects. A second stage comprises acts of searching for members of a predetermined specific-class of objects (such as humans in a certain pose) within the selected regions of the image using a cognitive swarm, detecting members of the specific-class of objects within the selected regions of the image, and outputting the locations of detected objects to an operator display and optionally to an automatic response system.

An object detection unit of a network camera detects an object including a human figure from image data. An event detection unit analyzes the image data based on a result of the object detection by the object detection unit. Based on a result of the image data analysis, an event detection unit determines to make identifiable to a client a human figure that should be monitored by the client and to apply masking to a human figure that should not be monitored by the client to make it unidentifiable to the client.

A method to equip a vehicle to perform object detection and tracking and a surround view camera system to perform the object detection and tracking involve two or more cameras arranged respectively at two or more locations of the vehicle. The cameras capture images within a field of view of the two or more cameras. A processing system obtains the images from the two or more cameras and performs image processing to detect and track objects in the field of view of the two or more cameras.

The invention provides an object detection and identification method, device, electronic equipment and a storage medium; the object detection and identification method based on multi-sensor fusion comprises the following steps: obtaining visual observation data of one or more objects according to image data of an image acquisition device; clustering laser data of a laser radar sensor, and obtaining position information of one or more objects to serve as the laser observation data; obtaining millimeter wave observation data of one or more objects according to millimeter wave data of a millimeter wave radar sensor; fusing and filtering the visual observation data, the laser observation data and the millimeter wave observation data of the same object according to time sequence so as to obtainthe fused observation data, and detecting and identifying the object according to the fusion data. The method and device can improve the object detection and identification reliability, stability andrecognition rate.

Methods, apparatuses, and computer-readable media are provided for maintaining blob trackers for video frames. For example, a first blob tracker maintained for a current video frame is identified. The first blob tracker is associated with a blob detected in one or more video frames. The blob includes pixels of at least a portion of a foreground object in the one or more video frames. It is determined that the first blob tracker is a first type of tracker. Trackers having the first type are associated with objects that have transitioned at least partially into a background model (referred to as sleeping objects and sleeping trackers). One or more interactions are identified between the first blob tracker and at least one other blob tracker. The at least one other blob tracker can be the first type of tracker or can be a second type of tracker that is not a sleeping tracker (the second type of tracker is not associated with an object that has transitioned at least partially into the background model. A characteristic of the first blob tracker can then be modified based on the identified one or more interactions. Modifying the characteristic of the first blob tracker can include transitioning the first blob tracker from the first type of tracker to the second type of tracker, updating an appearance model of the first blob tracker, and/or other suitable characteristic of the first blob tracker.

Computer-implemented methods for detecting objects within digital image data based on color transitions include: receiving or capturing a digital image depicting an object; sampling color information from a first plurality of pixels of the digital image; optionally sampling color information from a second plurality of pixels of the digital image; generating or receiving a representative background color profile based on the color information sampled from the first plurality of pixels; generating or receiving a representative foreground color profile based on the color information sampled from the second plurality of pixels and/or the first plurality of pixels; assigning each pixel a label; binarizing the digital image based on the labels; detecting contour(s) within the binarized digital image; and defining edge(s) of the object based on the detected contour(s). Corresponding systems and computer program products configured to perform the inventive methods are also described.

A method for moving objection detection in an image analysis system is provided. The method includes analyzing consecutive video frames from a single camera to extract box properties and exclude objects that are not of interest based upon the box properties. Motion and structure data are obtained for boxes not excluded. The motion and structure data are sent to a trained classifier. Moving object boxes are determined by the trained classifier. The moving object box identifications are provided to a vehicle system. The data sent to the classifier can consist of the motion and structure data, and no deep learning methods are applied to the video frame data. Driver assistance vehicle systems and autonomous driving systems are also provided based upon the moving object box detection.

Described is a system for stabilizing, detecting, and recognizing objects in video captured from a mobile platform. The system first receives a video (with a plurality of image frames) captured from a mobile platform. The video is stabilized by registering the image frames to a global coordinate system to generate stabilized image frames. A bio-inspired attention algorithm is applied to the stabilized image frames to produce a set of locations in the stabilized image frames that are salient points representative of an object of interest. An image chip is generated that surrounds each salient point. High-dimensional feature vectors are extracted from the image chip. The feature vectors are then classified as an object class. Thus, through classifying the feature vectors, an object of interest can be identified in the video as captured from the mobile platform.

Methods and systems for co-relating location and identity data available from Access Points (APs) and video surveillance systems are provided. According to one embodiment, data, including a unique identifier of an object and information regarding a first geo-position of the object, is received from an AP of a wireless network of a venue. A video feed captured by a camera system monitoring a portion of the venue and/or information regarding a second geo-position corresponding to the object are also received. The first and second geo-positions are then mapped to a common coordinate system. Based on the unique identifier, information regarding the object as reported by the AP and the camera system or derived therefrom are correlated. Finally, behavioral attributes of the correlated object are assessed based on one or a combination of actions of the object, the first and second geo-positions and the common coordinate system.

Disclosed are an object detection method and an object detection device. The object detection method comprises a step of obtaining plural detection results of a current frame according to plural object detection methods; a step of setting initial probabilities of the plural detection results of the current frame; a step of calculating a movement frequency distribution diagram representing movement frequencies of respective pixels in the current frame; a step of obtaining detection results of a previous frame; a step of updating the probabilities of the plural detection results of the current frame; and a step of determining a final list of detected objects based on the updated probabilities of the plural detection results of the current frame.

The present invention relates to a method for recognizing a transverse parking space 32 and auxilarily parking in the transverse parking space 23. A parking assistance system provides a method of identifying a transverse parking space 32. The method also supports parking in the parking space 32. For this purpose, a travel motion of a motor vehicle 10 is evaluated by means of the following method steps: The travel motion 38 is evaluated for straight travel and in particular for portions 44 during which the yaw angle of the motor vehicle 10 remains constant. At least one edge 70, 72 or one corner E1, E2 of a first object 34 delimiting the transverse parking space 32 is detected, and an edge 70, 72 or a corner E1, E2 of a second object 36 delimiting the transverse parking space 32 is detected.

The invention relates to a moving workpiece recognition method based on spatiotemporal contexts and a fully convolutional network, and belongs to the fields of digital image processing and object detection and recognition. According to the method, an object image database is utilized to train the fully convolutional neural network to obtain a classifier of a to-be-classified object; then a background difference method and a morphological method of digital image processing are utilized to obtain an initial position of the object in a first frame of a video sequence, an object tracking method of spatiotemporal context models is utilized to track the to-be-tracked object according to the initial position, and object tracking precision is verified through a precision graph; and finally, the trained classifier is utilized to carry out classification recognition on a tracking result, semantic-level segmentation is realized, and thus an object category is obtained. According to the method, the initial position of the moving object can be effectively and automatically acquired by using the background difference method and the morphological method of digital image processing, tracking and recognition for the moving workpiece on a conveyor belt can be realized, and an automation degree and an intelligence degree of an industrial robot are increased.