27 results about "Automatic target detection" patented technology

The present invention relates to the technical field of data simplification, in particular to a point cloud data simplification method for power line towers, and a system. The simplification method comprises: collecting point cloud data of power line towers, and carrying out data processing; performing data feature extraction on the processed data; classifying the point cloud data on the basis of the extracted features; and simplifying the classified point cloud data by means of using a point cloud simplification algorithm. The present invention introduces a deep learning technology to automatically learn and extract rich semantic features in point cloud data, thereby providing efficient feature representations for subsequent classification and simplification; by using extracted advanced semantic features and by means of a deep learning algorithm, the present invention can perform accurate classification and recognition on power line towers, thereby achieving automatic target detection and recognition; the present invention achieves intelligent simplification of point cloud data of power line towers by means of an intelligent merging strategy, so as to greatly reduce the data volume while reserving key features, thereby improving storage and transmission efficiency.

The invention relates to the technical field of computer vision and artificial intelligence, and provides a method and system for monitoring behaviors in an elevator. The method comprises the steps that an original frame image of an elevator is input into a target detection model, a target detection result is obtained, the target detection result comprises a target detection frame, a region of interest is determined based on the detection frame, the region of interest is input into a resolution reconstruction model, a high-resolution image is output, and the high-resolution image is the region of interest with the enhanced resolution. Fusing the region of interest with enhanced resolution with the original frame image to obtain a fused image, performing histogram equalization on the fused image to generate a target frame image, inputting the target frame image into a behavior recognition model to obtain a behavior recognition result, and obtaining a behavior monitoring result according to a target detection result and the behavior recognition result. Through automatic target detection and behavior recognition, an accurate, efficient and robust elevator monitoring solution can be provided.

The invention provides a target distance estimation method and system based on optical characteristics of a holder, and the method comprises the steps: carrying out the automatic target detection and tracking shooting of a to-be-detected target at a current visual angle, and obtaining a target video of the to-be-detected target; meanwhile, the cradle head posture and the camera focal length are collected in real time; performing target detection according to the target video to obtain key feature points of the target; performing distortion correction on the key feature points through image coordinate mapping to form a unified image coordinate system; combining the image coordinate system with the attitude information of the holder, and mapping the image coordinate system to a three-dimensional virtual space to construct a virtual coordinate system; and calculating the target distance of the to-be-measured target according to the geometric position of the to-be-measured target in the virtual coordinate system. According to the invention, the target distance can be accurately estimated in a complex scene and at different visual angles based on the holder system with a dynamic attitude adjustment function, and the method has the technical effects of real-time performance and robustness.

The invention discloses a target detection method and device based on a visual language large model. The method comprises the following steps: constructing a target detection data set for training; analyzing the image by using area detection and target center point regression, and extracting center point coordinates of a target; and taking the central point as a prompt, and inputting a segmentation model to obtain an initial segmentation mask of the target. The initial segmentation result is verified by means of a visual language large model, the initial result is subjected to reasoning analysis by means of a text prompt function guide model, and false positive and adhesion targets are recognized; and removing target adhesion and false positive objects by using an improved watershed method. Compared with the prior art, the scheme of the invention effectively alleviates the problem that the existing automatic target detection method fails under the complex conditions of fuzzy target gray distribution, adhesion and the like, and can improve the reliability and accuracy of target detection.

The invention is suitable for the field of computer vision and artificial intelligence, and provides an automatic target detection and tracking method and system based on deep learning, and the method comprises the steps: obtaining a video stream and a scene structure diagram, extracting a target position bounding box, a category label and an appearance feature vector frame by frame, and calculating a behavior parameter and a priority score; generating a detection frame interval and a tracking algorithm complexity parameter according to the priority score, and initializing a special tracker to realize automatic tracking; when the target is lost, generating a space-time probability graph based on historical data and the scene structure graph; determining retrieval parameters according to the priority scores, and performing directional detection on the screened space-time regions; and after the target is retrieved, the identity is associated and parameters are redistributed to recover tracking. According to the method, through differentiated resource allocation, accurate loss prediction and directional retrieval, tracking precision, efficiency and robustness are considered, and the problems of target tracking interruption, resource waste and the like in a complex scene are effectively solved.

A simulation radar fort device suitable for automatic target detection and strike teaching adopts double-steering-engine independent driving, connecting rod transmission and a gear reduction mechanism to realize horizontal scanning motion of an ultrasonic radar and response aiming motion of a fort. The radar scanning mechanism converts steering engine rotation into sector scanning through a connecting rod; the fort aiming mechanism adopts a reduction gear set to reduce the output angle of a steering engine, accurate angle adjustment is achieved, and quick response from radar discovery to fort aiming is achieved. Through combination of machinery and control, the device not only realizes spatial decoupling of scanning and aiming, but also can be used as a teaching platform to visually demonstrate angle transmission, error compensation and triangulation ranging principles in target positioning, and is suitable for teaching practices of courses such as automation and mechatronics.

The application is suitable for the field of computer vision and artificial intelligence, and provides an automatic target detection and tracking method and system based on deep learning, which comprises the following steps: acquiring a video stream and a scene structure graph, extracting a target position bounding box, a category label and an appearance feature vector frame by frame, calculating a behavior parameter and a priority score; generating a detection frame interval and a tracking algorithm complexity parameter according to the priority score, and initializing a special tracker to realize automatic tracking; when the target is lost, generating a space-time probability graph based on historical data and the scene structure graph; determining a search parameter according to the priority score, and performing directional detection on the screened space-time region; after the target is found back, associating the identity and reassigning the parameter to restore the tracking. Through differentiated resource allocation, accurate loss prediction and directional recovery, the application takes into account the tracking accuracy, efficiency and robustness, and effectively solves the problems of target tracking interruption and resource waste in a complex scene.

The application discloses a kind of underwater imaging quality evaluation methods based on UTOD, belong to underwater detection and imaging technical field, this method includes obtaining triangle target feature, and the image obtained is processed by human eye visual system HVS simulation module;Underwater triangle direction discrimination threshold UTOD model is constructed, and UTOD model is used to distinguish the direction of triangle target;Contrast-sharpness comprehensive index UICS is constructed, including underwater sharpness index and underwater contrast index;Based on four options forced choice 4AFC test, the direction discrimination correct probability corresponding to different UICS values is counted, and the relationship curve of UICS and correct probability is fitted by psychometric function Weibull;Triangle target of different size is repeatedly tested, and UTOD curve is generated.The application realizes automatic target detection and direction identification, and improves the accuracy, stability and objectivity of evaluation index.

The application discloses an unmanned aerial vehicle target detection optimization method and system based on privileged information knowledge migration, which comprises the following steps: analyzing a label file of an unmanned aerial vehicle target detection data set, extracting categories, boundary boxes and instance segmentation labels, generating a three-channel privileged information graph aligned with RGB images in the unmanned aerial vehicle target detection data set at a pixel level by using a privileged information processing method; splicing the RGB images and the three-channel privileged information graph in the channel dimension to form a six-channel image; constructing a teacher model and training the same; freezing the parameters of the trained teacher model, training a student model based on the RGB images by using a knowledge distillation loss function, and obtaining a trained student model; and automatically detecting a target of an RGB image to be detected by using the trained student model, and outputting a target category and a boundary box position. The application avoids dependence on complex multi-modal hardware, is simple and low in cost, and improves target detection performance.

The ice particle abrasive jet flow generation system of the present application comprises an ice particle preparation device, a high-pressure water jet device and an ice particle abrasive jet flow control device, the ice particle preparation device comprises an ice particle preparation cavity, a liquid nitrogen injector and a dye water injector, the ice particle abrasive jet flow control device comprises a dynamic monitoring cavity, a camera, a PTU controller and an interactive platform, the ice particle preparation device and the high-pressure water jet device respectively deliver ice particles and high-pressure jet water into the dynamic monitoring cavity, the ice particles and the high-pressure jet water can be mixed, forming ice particle abrasive jet water for deep sea mining, and the ice particles as abrasives can avoid causing pollution of seawater, in addition, after collecting image data through the camera, the optimized YoLo v11 algorithm is introduced for automatic target detection and capture of the ice particles in the abrasive jet flow in the dynamic monitoring cavity, and according to the development of the abrasive jet flow and the evolution process of the ice particles, the flow and pressure of the liquid nitrogen injector and the dye water injector are feedback adjusted to realize precise control of the ice particle abrasive jet flow.

This invention discloses a model-assisted semi-automatic target detection and annotation method for power line inspection, belonging to the fields of machine vision and power line inspection. The method includes: obtaining a basic annotation set using the power transmission and distribution images to be annotated; training a detector based on a reference annotation set, the basic annotation set, and the corresponding images to obtain a predicted annotation set for the images to be annotated; comparing the differences between the annotations in the basic annotation set and the predicted annotation set according to the intersection-union ratio (IUU) calculation results and target category determination, and classifying them into different annotation sets; sequentially determining whether each annotation in the annotation set is acceptable and correcting the basic annotation set; further training the detector until convergence; and updating the reference annotation set and the corresponding images. This invention combines model prediction with manual judgment to achieve efficient acquisition of high-quality target detection and annotation data.

The invention discloses a system for automatically detecting and tracking a target, and relates to the technical field of multi-target tracking. The system comprises a sensor interface module, a target identification module, a dynamic association gate module, a multi-algorithm association engine, a state estimation module and a track management module. A nearest neighbor (NN) algorithm, a global nearest neighbor (GNN) algorithm, a probability data association (PDA) algorithm and a joint probability data association (JPDA) algorithm are fused, dynamic tracking gate screening and Kalman filtering enhancement updating are combined, 2048 targets can be stably tracked by a single site at the same time, the tracking real-time performance, the anti-interference capability and the accuracy in a complex scene are remarkably improved, and the tracking efficiency is improved. The method is suitable for the fields of military monitoring, intelligent transportation, unmanned aerial vehicle clusters, spacecraft monitoring and the like.

The invention relates to the field of automatic target detection, and discloses a glass bead bubble detection counting system and method based on a solution scheme, and the system comprises a conveying device, a container tray, a visual detection system, a glass bead feeding device, a filtering separation device and a solution recovery device; a container tray is arranged on an original node on the conveying device and conveyed to a feeding node, and a glass bead feeding device feeds materials into the container tray; the loaded container tray is conveyed to a visual detection node by a conveying device, and enters a filtering and separating device node after being subjected to visual detection by a visual detection system; after the filtering and separating device node separates the solution and the glass beads in the container tray, the solution flows back to the solution recycling device, the container tray is conveyed back to the original node by the conveying device, and counting of bubbles of the glass beads is completed; according to the invention, automatic detection and counting of glass bead bubbles can be realized, and the detection efficiency and precision are improved.

This invention relates to the field of image data processing, specifically to an intelligent monitoring system for reservoir dam safety based on image recognition. The system includes: a data acquisition module for continuously acquiring video images and sensor time-series data at fixed intervals; a model training module for training a specialized AI algorithm model; a real-time inference module for automatic target detection; a data structuring module for outputting standardized recognition results; a data fusion and verification module for performing source fusion and cross-verification on data; a risk assessment module for analysis and outputting accurate risk assessment results; an early warning triggering module for triggering corresponding level early warnings based on the assessment results; a message push module for pushing messages to management personnel through multiple channels; and an event handling module for synchronously recording the entire process of early warning events, tracking the handling progress, and forming a closed-loop business process.

A system and method for operating a robotic system to register unrecognized objects is disclosed. The robotic system may use first image data representative of an unrecognized object to derive an initial minimum viable region (MVR). The robotic system may analyze second image data representative of the unrecognized object to detect a condition representative of an accuracy of the initial MVR. The robotic system may register the initial MVR or an adjustment thereof based on the detected condition.

Disclosed embodiments may include a method for automatic object detection, identification, and tracking. This may include receiving video data from one or more sensors, processing the video data to detect one or more objects using a first group of one or more machine learning models, tracking characteristics of the one or more objects over time, classifying, using a second group of one or more machine learning models, each of the one or more objects based on the tracked characteristics, comparing each of the classifications to a database of known objects, predicting an estimated identity of each of the one or more objects based on the comparison, and predicting a future path of motion for each of the one or more objects based on the estimated identity and the tracked characteristics.

The invention belongs to the technical field of automatic target detection, and relates to a real-time target detection method suitable for unconventional target detection, which comprises the following steps of: acquiring a frame of scene image from a detector, and selecting an interested target for tracking; training a target feature model; establishing a target feature library, and storing data of the target feature library into a memory; acquiring a frame of scene image from a detector, and executing target detection; establishing a binary image; performing region growth, expansion and clustering calculation on the image to determine a suspected target; and comparing the suspected target with a target feature library to determine a target detection result. According to the method, when the target detection task is executed, detection of any interested target category in a scene can be increased in time, real-time operation in embedded hardware can be achieved, rapid response and real-time processing of the detection task are achieved, and an accurate target detection result is provided.

The invention discloses an underwater imaging quality evaluation method based on UTOD, and belongs to the technical field of underwater detection and imaging, and the method comprises the steps: obtaining the features of a triangular target, and carrying out the processing of an obtained image through a human eye vision system HVS simulation module; constructing an underwater triangular direction discrimination threshold UTOD model, and discriminating the direction of the triangular target by using the UTOD model; constructing a contrast-definition comprehensive index UICS, wherein the UICS comprises an underwater definition index and an underwater contrast index; based on a four-option forced selection 4AFC test, counting direction discrimination correct probabilities corresponding to different UICS values, and fitting a relation curve of the UICS and the correct probabilities through a psychological measurement function Weibull; and repeatedly testing the triangular targets with different sizes to generate a UTOD curve. According to the invention, automatic target detection and direction identification are realized, and the accuracy, stability and objectivity of evaluation indexes are improved.

The application provides a cerebrospinal fluid cell microscopic image detection method and related device; multi-scale features of a cerebrospinal fluid cell microscopic image are extracted, and the multi-scale features are fused to obtain fused features; the fused features are aligned to obtain aligned features, the aligned features are subjected to frequency band modulation enhancement to obtain enhanced features, the enhanced features are fused to obtain unified features, and target enhanced features are obtained from the unified features according to a pre-set output size; detection is performed according to the target enhanced features, and a detection result is obtained; by introducing feature enhancement, different-scale effective information can be more fully fused, the features of a specified detection layer are subjected to targeted enhancement, the problems of missed detection and false detection that are prone to occur under a high-confidence requirement are improved, and the detection stability and reliability are improved; the cerebrospinal fluid cell morphological image input by a user can be subjected to automatic target detection, thereby reducing the burden of manual film reading and labeling and improving the efficiency of image analysis and result output.

The present invention provides a method, device, computing device and storage medium for automatic target detection and tracking. The method includes: obtaining in real time video images collected by a pan-tilt head; performing target detection on each frame of the video images to obtain video images containing detected targets; determining the detected targets as tracking targets, and generating a tracking template image by using the video images containing the detected targets; the tracking template image includes feature information of the tracking targets; and controlling the pan-tilt head to track the tracking targets according to the tracking template image and the video images obtained in real time, so that the tracking targets are located within the video images. With this solution, it is possible to implement automatic detection and tracking of targets by using a single pan-tilt head, and at the same time, the cost and the difficulty of deployment and debugging can be reduced.

The invention discloses a prior knowledge-based large-batch image data rapid interpretation method and system, and belongs to the field of computer vision and generative artificial intelligence, and the method comprises the steps: S1, constructing a task scene data set, training a CNN-DE algorithm through the data set, and obtaining a model for detecting a target type and a target position; s2, constructing an image-text data set by using the training set in the step S1, and training a fine-tuning LVLM model by using the image-text data set; s3, based on the data set in the step S1, combining the image and using a text to label the image to construct a multi-modal aligned data set, and using the data set to finely adjust CLIP to obtain an auxiliary interpretation model based on content retrieval; and S4, integrating the three models into a system, adjusting input and output of the system, constructing a human-computer interaction interface, and docking a database. The invention provides a multi-model and multi-mode collaborative image automatic target detection scheme based on prior knowledge, and has industrial application value and prospect.

The ice particle abrasive jet generation system comprises an ice particle preparation device, a high-pressure water jet device and an ice particle abrasive jet control device, the ice particle preparation device comprises an ice particle preparation cavity, a liquid nitrogen ejector and a dyeing water ejector, and the ice particle abrasive jet control device comprises a dynamic monitoring cavity, a camera, a PTU controller and an interaction platform. The ice particle preparation device and the high-pressure water jet device convey ice particles and high-pressure jet water into the dynamic monitoring cavity respectively, the ice particles and the high-pressure jet water can be post-mixed to form ice particle abrasive jet water for deep-sea mining, the ice particles serve as abrasives to avoid seawater pollution, and in addition, after image data are collected through a camera, the image data can be stored in the dynamic monitoring cavity. An optimized YoLo v11 algorithm is introduced to dynamically monitor automatic target detection and capture of ice particles in abrasive jet in a cavity, flow and pressure of a liquid nitrogen ejector and a dyeing water ejector are fed back and adjusted according to development of the abrasive jet and the evolution process of the ice particles, and accurate control over the ice particle abrasive jet is achieved.

The present invention discloses a method for multi-view feature extraction and discrimination of radar images, comprising the following steps: S1, a radar platform collects ground target image samples; S2, preprocesses the collected radar image samples; S3, constructs a multi-view image combination data set; S4, builds a multi-view image combination feature extraction network: the smallest unit of the picture is changed from a pixel to a block of a preset size through a patch division layer, and the pixel values in a block are combined into a vector; the generated vector will sequentially pass through three consecutive stages, namely stage 1, stage 2 and stage 3; S5, builds a multi-view image combination feature discrimination network; S6, inputs the data set obtained in S3 into the multi-view image combination feature extraction network and the multi-view image combination feature discrimination network for training, and uses the trained networks to identify unknown radar images. The present invention can effectively improve the accuracy of radar image classification and enhance the performance of the radar automatic target detection system by extracting and discriminating multi-view features.

The invention relates to an ammunition feed (1) for a gun barrel, in particular of an armoured vehicle, having at least one magazine and a rammer (2), the gun barrel having a loading space, it being possible for an ammunition (3) to be introduced from the magazine into the loading space by means of the rammer (2), and it being possible for an ammunition (3) to be led out of the loading space into the magazine by means of the rammer (2).An ammunition (3) can be guided out of a loading space particularly quickly and reliably by means of the ammunition feed (1), in that a pawl (4) is arranged movably on an end of the rammer (2) facing the cartridge, the pawl (4) being positionable in an engagement position (5) in order to guide the ammunition (3) out of the loading space into the magazine, the pawl (4) being arranged in a recess (6) of the ammunition (3) in the engagement position (5), the ammunition feed (1) being designed as an automatic ammunition feed (1), a control device being provided, the control device being connected in terms of control technology to at least one actuator of the rammer (2) and at least one actuator of the pawl (4), a type of ammunition being automatically selected on the basis of data of automatic target detection and / or automatic target recognition, by means of which a high probability of hitting and destruction of a target to be combated can be achieved.

Disclosed embodiments may include a method for automatic object detection, identification, and tracking. This may include receiving video data from one or more sensors, processing the video data to detect one or more objects using a first group of one or more machine learning models, tracking characteristics of the one or more objects over time, classifying, using a second group of one or more machine learning models, each of the one or more objects based on the tracked characteristics, comparing each of the classifications to a database of known objects, predicting an estimated identity of each of the one or more objects based on the comparison, and predicting a future path of motion for each of the one or more objects based on the estimated identity and the tracked characteristics.

This application discloses an automatic target detection method for live-line power distribution monitoring based on image recognition. The method includes: acquiring images of the work site and preprocessing them to obtain standardized images; extracting multi-scale feature maps using a shared backbone feature extraction network; detecting candidate regions for protective equipment and key human body points using a dual-branch network; constructing a human posture chain functionally associated with the protective equipment based on the key points; calculating an initial association score between the candidate regions and key points; analyzing the structural integrity of the protective equipment to obtain structural feature vectors; calculating a posture consistency index based on the posture chain; and finally fusing the above parameters to obtain functional association credibility, thereby determining the protective equipment wearing status. This method achieves accurate detection of the protective equipment wearing status, adapts to complex work scenarios, reduces false positives and false negatives, and improves the automation and intelligence level of safety monitoring for live-line power distribution operations.