Trajectories Of Objects Research Articles

Features of Bearing on Underwater Object Using Phase Information of a Differential Stereo Sensor

Introduction. Safety of navigation and development of underwater mineral deposits require the accurate detection of various underwater objects. The literature discusses the issues of tracking their motion and trajectory. Sonar methods are proposed to maintain high accuracy of underwater object positioning. High accuracy of the bearing of stereo sensors with an ultrashort base is noted. However, this equipment is sensitive to the sampling rate of the signals, which causes “sampling noise”. There are no publicly available publications dedicated to the solution to this problem. The presented study is designed to fill this gap. This work is aimed to study the possibility of obtaining data clarifying information about the bearing of underwater objects through using the phase information about echoed probing signals and an additional procedure for resampling the source data.Materials and Methods. The location of the object was determined using the experimental complex for studying hydroacoustic sensors created by V.A. Shirokov and V.N. Milich at the Udmurt Federal Research Center, the Ural Branch of the Russian Academy of Sciences. A stereo sensor with a small base (30 mm) was used compared to the distance to the object (≈800–900 mm). Digital filtering methods and mathematical apparatus of correlation analysis of return hydroacoustic signals obtained by the phase method were used for data processing.Results. The results of comparing two methods for determining the bearing on an object are presented: by the difference in the time of arrival of the pulse-leading edges and by the maximum of the cross-correlation function (CCF). The change in bearing as the object moves, is graphically shown. The use of the leading edge of the signal caused small outliers of values along the entire bearing curve (less than 0.12 rad). At the maximum CCF, emissions were recorded only in some areas, but they were quite significant (about 0.17 rad). It showed how to select points corresponding to a smoother and more valid object trajectory, and how to work with erroneous points. The presented method of error correction can be implemented programmatically. With a quasi-harmonious signal, rare measurements of the original signal are interpolated by frequent calculated values. Thanks to this virtual increase in the sampling rate (oversampling), intermediate indicators can be recorded in the digitized source data. Interpolation of the signal values by a cubic spline allowed us to obtain 20 points for 1 period of the signal instead of 5 points in the original version. In this case, the trajectory formed with the maximum CCF is more correct.Discussion and Conclusion. The direction-finding problem can be solved with the accuracy required for practical application. Taking into account the factor of smoothness and continuity of the object's trajectory makes it possible to qualitatively correct the selection of the maximum of the cross-correlation function of the stereo sensor signals. The proposed methods have great potential for the development of underwater vision systems.

MODEL OF RANDOM-LIKE PLANAR TRAJECTORIES WITH INTERSECTIONS

Background. Recently the task of detecting and identifying trajectories of objects whose genuine purposes are uncertain or strike threatening has become extremely important. The known approaches produce insufficiently smooth trajectories. Objective. The purpose of the paper is to build a model of generating random-like planar trajectories, which would have sufficiently smooth curves. A trajectory may have self-intersections and may intersect other trajectories. Methods. Preliminarily two starting points on a plane are generated. The distance and angle between these points are calculated, which then are successively updated to calculate new trajectory points using the polar coordinate system. A trajectory of points is generated using values of normally distributed random variables with zero mean and unit variance and four values of -uniformly distributed random variables. Results. The random-like trajectory generator has the same time complexity as its predecessors, including the direction randomization generator and its modifications. Exemplary trajectories appear very realistic. Self-intersections are important to manoeuvre and confuse the opponent side. The trajectory has four parameters to adjust its heading, scattering of points, and intensity of turns and twists. These parameters serve as magnitudes to amplify the respective properties. The highest influence has the angle-scattering parameter. Four simple conditions can be embedded to fit the trajectory within a rectangular domain. Conclusions. The suggested model should serve either for generating trajectory datasets to train manoeuvring-object detectors on them or for masking reconnaissance. The model allows balancing the trajectory smoothness and randomness.

Stable solution to the problem of autonomous navigation of moving objects on analytical trajectory intervals based on the results of inertial measurements

Provision of solution of the problem of autonomous navigation on a long time interval is considered. The main problem when using inertial navigation systems is caused by the accumulation of positioning and angular orientation errors. It is shown that this problem can be solved by combining methods of nonlinear filtering in processing the results of inertial measurements of platform-free inertial navigation systems and methods of reducing the dimensionality of the vector of navigation parameters on trajectory sections described by analytical models. Orthodromic (shortest) and circular intervals, most typical for program trajectories of transport objects – highways, railroads, airlines, etc., are considered as such trajectory sections. For circular intervals, analytical dependencies of the spatial coordinates of strapdown inertial navigation systems were obtained for the first time. These dependencies make it possible to reduce the dimension of the vector of navigation parameters and, as a consequence, to build a model of an autonomous observer of navigation parameters to use equations excluded from the general system of these parameters. It is shown that using stochastic nonlinear filtering methods it is possible to solve the problem of noise-resistant autonomous navigation on trajectory intervals described by analytical models. The proposed approach to solving the problem of autonomous navigation of moving objects makes it possible to significantly reduce computational costs in the practical implementation of algorithms for nonlinear filtering of current navigation parameters in comparison with traditional models of strapdown inertial navigation systems. The effectiveness of the proposed approach is illustrated with a numerical example. The results obtained are useful in developing navigation support for various transport objects moving along program trajectories, for example, aircraft for various purposes, railway, road, river transport, etc.

Towing an Object With a Rover

Abstract The present study investigates the problem of towing an object that is lying on a surface in a given workspace and the applicability to a planetary rover with four steering wheels. A quasi-static method has been introduced and used for path planning and for the synthesis of both object and rover trajectories. The rover uses a tether as the towing medium which is modeled as an elastic unilateral constraint. Moreover, a kinematic model of the rover which includes steering asymmetrical joint limits is taken into account. The dynamics model of the overall system is then derived and a sensitivity analysis is performed over a finite number of different trajectories, in order to evaluate the quasi-static assumption, the effects of the model, and the influence of the elastic constraint. Finally, experiments have been performed using the novel Archimede rover prototype and compared with dynamics simulations; the remarkable adherence shown with the model validates the overall approach.

DMOT-SLAM: visual SLAM in dynamic environments with moving object tracking

Visual simultaneous localization and mapping (SLAM) in dynamic environments has received significant attention in recent years, and accurate segmentation of real dynamic objects is the key to enhancing the accuracy of pose estimation in such environments. In this study, we propose a visual SLAM approach based on ORB-SLAM3, namely dynamic multiple object tracking SLAM (DMOT-SLAM), which can accurately estimate the camera’s pose in dynamic environments while tracking the trajectories of moving objects. We introduce a spatial point correlation constraint and combine it with instance segmentation and epipolar constraint to identify dynamic objects. We integrate the proposed motion check method into DeepSort, an object tracking algorithm, to facilitate inter-frame tracking of dynamic objects. This integration not only enhances the stability of dynamic features detection but also enables the estimation of global motion trajectories for dynamic objects and the construction of object-level semi-dense semantic maps. We evaluate our approach on the public TUM, Bonn, and KITTI dataset, and the results show that our approach has a significant improvement over ORB-SLAM3 in dynamic scenes and performs better compared to other state-of-the-art SLAM approaches. Moreover, experiments in real-world scenarios further substantiate the effectiveness of our approach.

Интеллектуальная система управления траекторным движением морского подвижного объекта и классификации объектов

Актуальность вопроса разработки интеллектуальных систем управления и классификации динамических объектов продиктована тем, что в области применения систем искусственного интеллекта появился большой класс сложных объектов требующих управления в реальном времени, функционирующих в проблемной (неопределенной) среде. Например, морские подвижные объекты (в том числе беспилотные), беспилотные летательные аппараты палубного базирования и т.п. Построение систем управления указанными объектами с использованием классической теории управления не позволяет получить требуемое качество управления в реальном времени. В связи с этим в работе приведены основные принципы построения интеллектуальных систем управления сложными динамическими процессами и классификации наблюдаемых им объектов с использованием нечёткого способа представления знаний. Рассмотрены вопросы построения и проблемы реализация логико-лингвистических (нечетких) моделей принятия управленческих решений в системах искусственного интеллекта, функционирующих в реальном масштабе времени. Проведён анализ современных способов обработки нечёткой информации и средств реализации нечётких алгоритмов управления. Отмечено, что несоответствие структур современных вычислительных устройств параллельному характеру обработки нечёткой информации в процессе реализации нечётких алгоритмов управления не позволяет в полной мере использовать возможности построения и реализации логико-лингвистических моделей управления сложными процессами и классификации динамических объектов в реальном масштабе времени. Предложены обобщённые структурные элементы однородных процессоров, на базе которых созданы трёхмерные ассоциативные модули специализированного сопроцессора поразрядно-параллельного принципа действия, служащего для обработки знаний в процессе принятия решений, а также структуру интеллектуальной системы управления траекторным движением морского подвижного объекта и классификации объектов на основе нечеткой базы знаний. Annotation. The relevance of the issue of developing intelligent control systems and classification of dynamic objects is dictated by the fact that in the field of application of artificial intelligence systems a large class of complex objects that require real-time control and operate in a problematic (uncertain) environment has appeared. For example, marine mobile objects (including unmanned ones), deck-based unmanned aerial vehicles, etc. The construction of control systems for these objects using classical control theory does not allow obtaining the required quality of control in real time. In this regard, the paper presents the basic principles of constructing intelligent control systems for complex dynamic processes and classifying objects observed by them using a fuzzy method of representing knowledge. The issues of construction and problems of implementation of logical-linguistic (fuzzy) models of management decision-making in artificial intelligence systems operating in real time are considered. An analysis of modern methods of processing fuzzy information and means of implementing fuzzy control algorithms has been carried out. It is noted that the discrepancy between the structures of modern computing devices and the parallel nature of processing fuzzy information in the process of implementing fuzzy control algorithms does not allow full use of the possibilities of constructing and implementing logical-linguistic models for controlling complex processes and classifying dynamic objects in real time. Generalized structural elements of homogeneous processors are proposed, on the basis of which three-dimensional associative modules of a specialized coprocessor of a bit-parallel operating principle are created, which serves for processing knowledge in the decision-making process, as well as the structure of an intelligent system for controlling the trajectory movement of a marine mobile object and classifying objects based on a fuzzy knowledge base.

Clairvoyance: Vision-impaired friendly assistive mobile device

Our design is a wearable device that assists the visually impaired to move. Our goal is to enable visually impaired people to travel alone after wearing our designs, helping them reduce the risks they may face when walking out alone. After several iterations of design ideas, our final design mainly relies on two webcams, Jetson TX2 Development board, and six vibrators. These components are installed on a sports vest and a belt. We decided to use visual systems and GPS to predict the trajectory of people and objects, and help users identify the direction of obstacles and achieve the purpose of obstacle avoidance through vibrators at different positions.

Efficient and robust 3D reconstruction: Absolute phase extraction algorithm based on three gray images

The structured light three-dimensional (3D) measurement technology based on optical metrics has gained increasing applications, and there is a growing demand for capturing transient trajectory changes in dynamic scenes. Conventional temporal phase unwrapping (TPU) approaches often require multi-frequency fringe patterns for application. However, the inevitable conflict arises between the number of projected fringe patterns and the measurement speed. This paper proposes a highly accurate and robust 3D measurement method using only 3-step phase-shifting fringe patterns. To precisely reconstruct the 3D shape using the fewest fringe patterns, it is essential to extract as much information as possible from the available patterns. Therefore, we calculate the wrapped phase distribution using 3-step phase-shifting fringe patterns to ensure accuracy. Simultaneously, the trapezoidal phase-shifting method is employed to calculate the intensity ratio distribution. Through the proposed region-based indexing method, the intensity ratio map is transformed into an unwrapped phase map. Additionally, a novel self-correction method is introduced to correct the non-linear errors in the unwrapped phase map, which is then utilized for the calculation of the fringe order required for phase unwrapping. This high-speed measurement method utilizes only 3-step phase-shifting fringe patterns while retaining the advantages of TPU, meeting the requirements of high-speed and full-field measurements. Experimental results on both static and dynamic scenes validate the effectiveness and versatility of the proposed method, demonstrating its flexibility and simplicity in application. In our 3D measurement system, the method achieves 3D reconstruction of the dynamic object's trajectory at a measurement speed of 107 frames per second.

Research of the Influence of Lateral Inflow Angles on the Cavitation Flow and Movement Characteristics of Underwater Moving Objects

This study examined the multi-phase flow field for a single object and two parallel/series objects under different incoming angles of lateral flow. The volume of fluid model, the Sauer–Schnerr cavitation model, and the six degrees of freedom (DOF) method were adopted to consider simulations of multi-phase flow, phase change, and object movement, respectively. The results show that, for a single object, the degree of asymmetry in the cavity profile depends on the component (the z-component) of the lateral inflow velocity in the direction perpendicular to the initial velocity of the object. As this component increases, the asymmetry of the cavity increases. The cavity length is related to the relative axial speed between the object and the water. For parallel objects, the cavity asymmetry is determined by the superimposed influence of the z-component of the lateral incoming speed and the high-pressure zone induced by the nearby object. The object located downstream relative to the lateral flow has a stronger cavity asymmetry than that of the upstream object, and the trajectory of the downstream object is more easily deviated than that of the upstream object. For tandem objects, with the increase in the lateral incoming angle, the supercavity length increases after the rear object enters into the front cavity. With the increase in the z-component of the lateral flow velocity, the deviation speed increases.

Braneworld black holes with ModMax electrodynamics on the brane

In this study, we investigate ModMax electrodynamics localized within the Randall–Sundrum II and Dvali–Gabadadze–Porrati branes, deriving corresponding 3-brane spacetime solutions that conform to the effective Einstein equations in each specific scenario. We construct solutions for charged black holes within the effective Einstein equation framework in both braneworld scenarios. The examination explores the trajectories of charged objects in this spacetime, underscoring the significance of the nonlinear parameter. Our analysis uncovers the fact that, similar to the prior ModMax black hole investigation, the nonlinear parameter plays a pivotal role in suppressing the effective charge of the black hole, due to its definite positive value.

Cueing Flight Object Trajectory and Safety Prediction Based on SLAM Technology

With the rapid development of artificial intelligence and robot technology, SLAM technology, as a key component, has been paid more and more attention. SLAM technology enables robots to autonomously navigate, build maps, and achieve accurate positioning in unknown environments, providing strong support for the autonomy and intelligence of robots and unmanned vehicles. In this paper, the position prediction method of flying object based on SLAM technology and the application of EvolveGCN model in behavior prediction are introduced. First, through the fusion of SLAM technology and liDAR data, we can accurately predict the position and movement trajectory of flying objects, thereby improving the safety and efficiency of the system. Secondly, with the EvolveGCN model, we are able to capture dynamic changes in the environment and achieve accurate predictions of the behavior of flying objects. Through experimental verification, the prediction accuracy of our method has been significantly improved in both simulation and real environment, which indicates the feasibility and effectiveness of the method in practical application, and provides an important reference and technical support for the development of autonomous navigation, aerial surveillance and other fields.

Paediatric Activity Anomaly Detection using CNN-XGBoost for Early Intervention and Patient Safety

The development of a robust anomaly detection system capable of distinguishing between normal and abnormal activities in paediatric care settings is of paramount importance for early intervention and patient safety. This endeavor presents substantial challenges due to the unique characteristics of paediatric activities, making it necessary to employ a combination of Convolutional Neural Networks (CNNs) and XGBoost to effectively address these complexities. The initial step involves the assembly of a meticulously annotated dataset comprising video sequences encompassing a broad spectrum of paediatric activities, both normal and abnormal, with a strong emphasis on diversity and representativeness. This dataset is subjected to rigorous preprocessing to ensure consistency and quality. This research employs a Kalman filter as a pre-processing step. This filter helps to reduce noise and enhance the quality of the video data. By smoothing and stabilizing the trajectories of objects or subjects within the video frames, the Kalman filter provides a cleaner input for subsequent stages of the process. CNN employed for feature extraction from video frames, capturing both spatial and temporal cues. To account for the temporal dimension inherent in video data, by ensuring that the model can capture nuanced patterns in paediatric activities. The accuracy of the proposed method is 88%. The extracted features or sequences are input into an XGBoost classifier, designed as a binary classification task to differentiate between normal and abnormal activities. Robust model evaluation is conducted on validation and test datasets using appropriate performance metrics, with continuous feedback loops established with healthcare professionals and caregivers to fine-tune and optimize the model's performance.

Human-in-the-Loop: Visual Analytics for Building Models Recognizing Behavioral Patterns in Time Series.

Detecting complex behavioral patterns in temporal data, such as moving object trajectories, often relies on precise formal specifications derived from vague domain concepts. However, such methods are sensitive to noise and minor fluctuations, leading to missed pattern occurrences. Conversely, machine learning (ML) approaches require abundant labeled examples, posing practical challenges. Our visual analytics approach enables domain experts to derive, test, and combine interval-based features to discriminate patterns and generate training data for ML algorithms. Visual aids enhance recognition and characterization of expected patterns and discovery of unexpected ones. Case studies demonstrate feasibility and effectiveness of the approach, which offers a novel framework for integrating human expertise and analytical reasoning with ML techniques, advancing data analytics.

Object Detection and Tracking with YOLO and the Sliding Innovation Filter.

Object detection and tracking are pivotal tasks in machine learning, particularly within the domain of computer vision technologies. Despite significant advancements in object detection frameworks, challenges persist in real-world tracking scenarios, including object interactions, occlusions, and background interference. Many algorithms have been proposed to carry out such tasks; however, most struggle to perform well in the face of disturbances and uncertain environments. This research proposes a novel approach by integrating the You Only Look Once (YOLO) architecture for object detection with a robust filter for target tracking, addressing issues of disturbances and uncertainties. The YOLO architecture, known for its real-time object detection capabilities, is employed for initial object detection and centroid location. In combination with the detection framework, the sliding innovation filter, a novel robust filter, is implemented and postulated to improve tracking reliability in the face of disturbances. Specifically, the sliding innovation filter is implemented to enhance tracking performance by estimating the optimal centroid location in each frame and updating the object's trajectory. Target tracking traditionally relies on estimation theory techniques like the Kalman filter, and the sliding innovation filter is introduced as a robust alternative particularly suitable for scenarios where a priori information about system dynamics and noise is limited. Experimental simulations in a surveillance scenario demonstrate that the sliding innovation filter-based tracking approach outperforms existing Kalman-based methods, especially in the presence of disturbances. In all, this research contributes a practical and effective approach to object detection and tracking, addressing challenges in real-world, dynamic environments. The comparative analysis with traditional filters provides practical insights, laying the groundwork for future work aimed at advancing multi-object detection and tracking capabilities in diverse applications.

Culturology of Journalism: Academic and Pragmatic Dimensions

The article attempts to represent the culturology of journalism as a relatively new area of theoretical research and an autonomous academic discipline. From a scientific point of view, the basis for the separation of this direction should be found in the sectoral culturology forming process, which includes a number of disciplines such as, for example, cultural studies of religion, education, politics, everyday life and, in particular, media space; the article contains detailed references to the views and proposals of the scholars in cultural studies of art. In each case, specific object, subject, institutional status and development trajectories are determined, that directly concerns the journalism theory. At the same time, the author aims to identify those characteristics of the discipline that enable it to develop fruitfully in interaction with the real media industry and, most importantly and urgently, the professionals' training system. The importance of this area in the academic dimension increases if the real problems and tasks of mass information practice are reflected in research activities and ways to enhance it are identified, recommendations are made regarding improving and enriching the professional qualifications of media organizations employees. This is the principal methodological position of the author. The field of university journalism education deserves special attention, since cultural knowledge and skills should be among the core competencies of graduates and find application in their professional and creative activities. The article presents judgments and pedagogical approaches to solving this problem, which are found in foreign and domestic literature. The author reveals in detail the experience of St. Petersburg State University, where for decades the journalism culturology has been purposefully developing in research and educational projects of various forms.

Multi-Scene Generalized Trajectory Global Graph Solver with Composite Nodes for Multiple Object Tracking

The global multi-object tracking (MOT) system can consider interaction, occlusion, and other ``visual blur'' scenarios to ensure effective object tracking in long videos. Among them, graph-based tracking-by-detection paradigms achieve surprising performance. However, their fully-connected nature poses storage space requirements that challenge algorithm handling long videos. Currently, commonly used methods are still generated trajectories by building one-forward associations across frames. Such matches produced under the guidance of first-order similarity information may not be optimal from a longer-time perspective. Moreover, they often lack an end-to-end scheme for correcting mismatches. This paper proposes the Composite Node Message Passing Network (CoNo-Link), a multi-scene generalized framework for modeling ultra-long frames information for association. CoNo-Link's solution is a low-storage overhead method for building constrained connected graphs. In addition to the previous method of treating objects as nodes, the network innovatively treats object trajectories as nodes for information interaction, improving the graph neural network's feature representation capability. Specifically, we formulate the graph-building problem as a top-k selection task for some reliable objects or trajectories. Our model can learn better predictions on longer-time scales by adding composite nodes. As a result, our method outperforms the state-of-the-art in several commonly used datasets.

QDETRv: Query-Guided DETR for One-Shot Object Localization in Videos

In this work, we study one-shot video object localization problem that aims to localize instances of unseen objects in the target video using a single query image of the object. Toward addressing this challenging problem, we extend a popular and successful object detection method, namely DETR (Detection Transformer), and introduce a novel approach –query-guided detection transformer for videos (QDETRv). A distinctive feature of QDETRv is its capacity to exploit information from the query image and spatio-temporal context of the target video, which significantly aids in precisely pinpointing the desired object in the video. We incorporate cross-attention mechanisms that capture temporal relationships across adjacent frames to handle the dynamic context in videos effectively. Further, to ensure strong initialization for QDETRv, we also introduce a novel unsupervised pretraining technique tailored to videos. This involves training our model on synthetic object trajectories with an analogous objective as the query-guided localization task. During this pretraining phase, we incorporate recurrent object queries and loss functions that encourage accurate patch feature reconstruction. These additions enable better temporal understanding and robust representation learning. Our experiments show that the proposed model significantly outperforms the competitive baselines on two public benchmarks, VidOR and ImageNet-VidVRD, extended for one-shot open-set localization tasks.

Maximum Principle in Autonomous Multi-Object Safe Trajectory Optimization

The following article presents the task of optimizing the control of an autonomous object within a group of other passing objects using Pontryagin’s bounded maximum principle. The basis of this principle is a multidimensional nonlinear model of the control process, with state constraints reflecting the motion of passing objects. The analytical synthesis of optimal multi-object control became the basis for the algorithm for determining the optimal and safe object trajectory. Simulation tests of the algorithm on the example of real navigation situations with various numbers of objects illustrate their safe trajectories in changing environmental conditions. The optimal object trajectory obtained using Pontryagin’s maximum principle was compared with the trajectory calculated using the Bellman dynamic programming method. The analysis of the research allowed for the formulation of valuable conclusions and a plan for further research in the field of autonomous vehicle control optimization. The maximum principle algorithm allows one to take into account a larger number of objects whose data are derived from ARPA anti-collision radar systems.

Efficient LiDAR-Trajectory Affinity Model for Autonomous Vehicle Orchestration

Computation and memory resource management strategies are the backbone of continuous object tracking in intelligent vehicle orchestration. Multi-object tracking generates enormous measurements of targets and extended object positions using light detection and ranging (Lidar) sensors. Designing an adequate object-tracking system is a global challenge because of dynamic object detection and data association uncertainties during scene understanding. In this regard, we develop an intelligent multi-objective tracking (IMOT) system with a novel measurement model, called the box data association inflate (BDAI) model, to assess each target’s object state and trajectory without noise by using the Bayesian approach. The box object filter method filters ambiguous detection responses during data association. The theoretical proof of the box object filter is derived based on binomial expansion. Prognosticating a lower-dimension object than the original point object reduces the computational complexity of vehicle orchestration. Two datasets (NuScenes dataset and our lab dataset) are considered during the simulations, and our approach measures the kinematic states adequately with reduced computation complexity compared to state-of-the-art methods. The simulation outcomes show that our proposed method is effective and works well to detect and track objects. The NuScenes dataset contains 28130 samples for training, 6019 examples for validation and 6008 samples for testing. IMOT achieves 58.09% tracking accuracy and 71% mAP with 5 ms pre-processing time. The Jetson Xavier NX consumes 49.63% GPU and 9.37% average power and exhibits 25.32 ms latency compared to other approaches. Our system trains a single pair frame in 169.71 ms with affinity estimation time of 12.19 ms, track association time of 0.19 ms and mATE of 0.245 compared to state-of-the-art approaches

Contactless weighing method based on deep learning and acoustic levitation

Abstract Acoustic weighing is a promising contactless method for screening the mass of micro-nano objects as it avoids contact contamination and losses. Existing acoustic weighing methods determine the mass of an object by detecting its oscillation trajectory with a laser sensor. However, this method suffers from several limitations, such as short measurement distance, poor accuracy in measuring transparent objects, and inducing damage to photosensitive samples. To solve these issues, this work proposes a contactless weighing method based on location-aware neural network (LANet) and acoustic levitation. The proposed LANet is a deep learning-based image processing method that detects object bit oscillation trajectories completely contactless, regardless of the color, shape, and oscillation distance of the levitated object. We employ a cross-stage aggregation module and cross-mixed feature pyramid strategy to build LANet network depth for enhanced feature extraction. In addition, to create a contactless environment, we built an acoustic levitation system, which drives the oscillation of objects. Finally, we verified the accuracy and effectiveness of the method. The results show that the proposed network can accurately detect the oscillation trajectories of various objects with high detection performance, even for small objects in low-contrast backgrounds. Meanwhile, the proposed method can accurately measure the mass of objects with a percentage error of no more than 7.83%.