Edge neuro-statistical learning for event-based visual motion detection and tracking in roadside safety systems

In order to make the service robot carry out dynamic path planning according to the movement state of pedestrians and improve the ability of collaboration with humans, this paper proposes a pedestrian detection and tracking algorithm based on 2D lidar. Pedestrian detection based on 2D lidar is mainly done by detecting the legs of pedestrians. In the laser data preprocessing stage, in order to reduce the influence of lidar point cloud noise, Gaussian filtering is used to filter the original point cloud data. In order to reduce the influence of the environment itself on the detection accuracy, an algorithm for projecting the laser data with the static map is proposed to remove the static background data. In the clustering segmentation stage, this paper uses Euclidean clustering to segment the laser point cloud clustering. In the feature extraction stage, in order to improve the detection accuracy of pedestrians' legs, the number of statistical features is increased on the basis of extracting geometric features; In order to improve the generalization ability of the model, and to solve the problem of scale inconsistency of pedestrian leg feature description, the extracted pedestrian features are normalized. This article uses SVM to build a human leg classifier. In the pedestrian tracking phase, this paper proposes a multi-person tracking algorithm based on Kalman filtering. The experimental results show that the proposed scheme can obtain better pedestrian detection and tracking effects, and the real-time performance is higher.

This paper describes results of experiments with camera setup, calibration and image processing algorithms for automatic detection and tracking of pedestrians and vehicles. The aim of the MOBIS project was to develop a method of assessing safety of unsignalised pedestrian crossings. Correct detection and tracking proved to be more difficult in the case of pedestrians than vehicles due to variability in people’s appearance, movement in groups and poor visibility in bad weather. Application of cameras with built-in pedestrian tracking programs was successful only in very good visibility conditions, so a computationally efficient PC algorithm providing a high pedestrian detection rate was used instead. The paper presents comparison of results obtained using different image processing methods as well as selected problems of pedestrian tracking. Statistical analysis of pedestrian behaviour with and without vehicles present is also shown. The proposed approach seems to be accurate enough for the purpose of assessing pedestrian safety.

High-quality pedestrian tracking algorithms are the basis for solving many problems including video surveillance, action recognition, and crowd behaviour analysis. The problem of pedestrian target detection and tracking is to identify and track the trajectories of different targets in video sequences. It can be divided into single target and tracking, multi-target and tracking, pedestrian re-identification, and so on. In recent years, with improvement of benchmarking, pedestrian object tracking methods have also developed from traditional feature learning to neural network modelling. This paper mainly summarizes and analyses the latest proposed deep learning algorithm models for pedestrian detection and tracking and the corresponding improvement for existing problems. In addition, possible future research directions are also discussed at the end.

This paper addresses visual motion tracking by a connectionist method, and aims at showing how the flexibility and the generalization power of neural networks can enhance a tracking system's adaptiveness and effectiveness. The simple principle of operation widens the range of applicability. A set of tracking structures that exhibit increasing levels of integration and efficiency are described. We also show how multinetwork architectures for estimate averaging may greatly increase tracking stability. The validity of the basic mechanism was assessed on a simple domain; however, a specific difficult testbed made it possible to verify the effectiveness of the method.

Pedestrian detection is one of the significant task in any intelligent systems involving video surveillance, since it provides essential information regarding the semantic behavior of pedestrians from video footages. Pedestrian detection along with tracking serves as an obvious extension to automotive applications in design and improvement of safety systems. However, various challenges arise while designing a system for detection and tracking of pedestrians such as different styles of clothing, non-linear random motion of pedestrians, occlusions between pedestrians and surroundings. Particle filtering algorithm is best suited to overcome these types of difficulties. Existing approaches such as Kalman filtering technique are also being implemented in this paper in order to compare the results and further prove that the Particle filtering dominates over the existing approaches.

A review on the use of top-view surveillance videos for pedestrian detection, tracking and behavior recognition across public spaces.

In the field of computer vision, pedestrian detection is a significant research area that is frequently applied in areas like intelligent driving and video surveillance. The reliability and accuracy of subsequent tasks like pedestrian tracking and behavior analysis are directly impacted by the accuracy and effectiveness of pedestrian detection. However, there are still a number of difficulties with pedestrian detection tasks because of the variety of factors, including the shape, scale, and posture of pedestrians, as well as background interference and changes in lighting. Based on research on pertinent domestic and international literature, this article examines the key techniques for identifying pedestrians. Introduce frequently used datasets for pedestrian detection first. The model can then be split into two categories based on its training process: one-stage detection algorithms and two-stage detection algorithms. For each stage type, this study offers a thorough introduction to representative algorithms. A summary of its current state of development and future prospects was then given.

Visual tracking is a fundamental technology in computer vision, with wide-ranging applications in fields such as robotics, autonomous driving, augmented reality, and security systems. A key challenge in these domains is ensuring stable object tracking in dynamic and complex environments. Tracking objects in real-world settings presents difficulties like occlusion, lighting changes, and sudden object movements, all of which demand the development of robust algorithms that can handle such variability effectively. This research introduces a novel single object visual tracking technique that leverages the YOLO (You Only Look Once) object detection model, combined with an adaptive particle filter. By integrating YOLO's precise object detection capabilities with the continuous state estimation offered by particle filters, this approach achieves stable object tracking, even in challenging environments. At the heart of the proposed algorithm is a dynamic fusion of YOLO detection results with image feature-based particle weight updates. When YOLO successfully detects an object, the position and size data of the object are used to adjust the state and weights of the particles, resulting in improved tracking accuracy and quick recovery from events such as sudden movements or occlusions. Conversely, in situations where YOLO detection is unavailable, the algorithm seamlessly transitions to a pure particle filter mode. In this fallback mode, visual features, such as image color histograms, are used to update particle weights, ensuring continuous and robust tracking even during temporary detection failures. The object’s dynamic state is modeled using an 8-dimensional vector, which includes information about position, velocity, size, and size-change rate. This comprehensive state representation enables precise tracking of various object motions and size transformations. Additionally, a Bhattacharyya distance-based weight calculation method is employed to assess particle similarity more effectively. Experimental results demonstrate that the proposed method excels in various challenging scenarios, including those with sudden object movements, partial occlusions, and lighting variations. The algorithm consistently delivers stable tracking, underscoring its robustness and adaptability. The contributions of this research have potential applications in fields such as autonomous driving, SLAM (Simultaneous Localization and Mapping), and augmented reality. Specifically, this technique could enhance pedestrian and vehicle tracking in autonomous driving systems and improve the accuracy of dynamic object tracking and environmental mapping in SLAM. It is also likely to play a key role in real-world applications like person tracking in security systems, player motion analysis in sports, and focused tracking in medical imaging. Future research directions include expanding the method to handle multi-object tracking and optimizing its real-time performance, broadening its applicability. Additionally, further improvements in tracking accuracy and robustness may be achieved by incorporating deep learning-based feature extraction and integrating data from various sensors.

Abstract. Pedestrian detection and tracking remains a highlight research topic due to its paramount importance in the fields of video surveillance, human-machine interaction, and tracking analysis. At present time, pedestrian detection is still an open problem because of many challenges of image representation in the outdoor and indoor scenes. In recent years, deep learning, in particular Convolutional Neural Networks (CNNs) became the state-of-the-art in terms of accuracy in many computer vision tasks. The unsupervised learning of CNNs is still an open issue. In this paper, we study a matter of feature extraction using a special activation function. Most of CNNs share the same architecture, when each convolutional layer is followed by a nonlinear activation layer. The activation function Rectified Linear Unit (ReLU) is the most widely used as a fast alternative to sigmoid function. We propose a bounded randomized leaky ReLU working in such manner that the angle of linear part with the highest input values is tuned during learning stage, and this linear part can be directed not only upward but also downward using a variable bias for its starting point. The bounded randomized leaky ReLU was tested on Caltech Pedestrian Dataset with promising results.

This paper presents a novel method for simultaneous pedestrian detection and tracking in image sequences. The motion detection and tracking are addressed in a common framework that employs a geometric active contour model, in which an evolving curve was formulated by the level set method. The pedestrian detection module is started after the curve evolution, which stops the propagating contour on the object boundary. The HOG features are extracted on a set of points located on a narrow band around the zero level set, and serve as cues for pedestrian detection. Experimental results demonstrate that the proposed HOG descriptor provides more information representing pedestrians' distinguishing shape features than the original one.

Aiming at the problem of low pedestrian tracking accuracy caused by illumination and occlusion in intelligent surveillance videos, a pedestrian tracking algorithm based on Work-Yolo detection combined with DeepSORT is proposed. To improve the accuracy of the detector, the attention module CBAM is used to fuse with the Backbone and Neck parts of Yolov5s network for enhancing pedestrian features. The BiAdd structure is used to fuse features of different scales about BiFPN, Dilated ConV is proposed to reduce the number of model parameters and extract better shallow features. Work-Yolo head decoupled head separates prediction classification and regression tasks, solving the problem of missed detection when pedestrians are obscured. The Lite-shufflenet lightweight structure is proposed to extract appearance features, and retrain the pedestrian re-identification dataset to reduce the identity switching caused by pedestrian occlusion. Pedestrian detection experiments are conducted on 3247 intersection pedestrian datasets, and the final detection accuracy rate was 94.2% and the recall rate was 90.6%. The video taken at the indoor entrance and exit of the scene, four videos are randomly selected for multi-target tracking experiments, the MOTA is improved by 10%, and the detection speed reaches 15fps, which meets the requirements of industrial applications.

Dancing robot should plan a dynamic path based on the motion of pedestrians to ensure the safety of others, so a pedestrian recognition and tracking method based on laser scanner is proposed. At first, the data collected by the laser is preprocessed. To segment the environment rapidly, the DBSCAN algorithm is optimized to cluster the data. Then a pedestrian recognition method is proposed by using the distance from cluster to laser scanner and pedestrian width range. At last, the laser data is converted into video data frames. So the particle filter algorithm is adopted to track the moving pedestrian. Experimental results show that the proposed method has a good real-time performance and high application value.

Pedestrian detection and tracking in infrared imagery using shape and appearance

Pedestrian detection and tracking is a critical task in the area of smart building surveillance. Due to advancements in sensors, the architects concentrate in construction of smart buildings. Pedestrian detection in smart building is greatly challenged by the image noises by various external environmental parameters. Traditional filter-based techniques for image classification like histogram of oriented gradients filters and machine learning algorithms suffer to perform well for huge volume of pedestrian input images. The advancements in deep learning algorithms perform exponentially good in handling the huge volume of image data. The current study proposes a pedestrian detection model based on deep convolution neural network (CNN) for classification of pedestrians from the input images. Proposed optimized version of VGG-16 architecture is evaluated for pedestrian detection on the INRIA benchmarking dataset consisting of 227 × 227 pixel images. The proposed model achieves an accuracy of 98.5%. It was found that proposed model performs better than the other pretrained CNN architectures and other machine learning models. Pedestrians are reasonably detected and the performance of the proposed algorithm is validated.

Target detection and tracking technology is particularly important in the field of computer vision. With the needs of social production and the development of science and technology, the research of video pedestrian detection and tracking technology based on mobile platforms also needs to be put on the agenda. In this paper, the mobile robot is based on the Gmapping method to map the unknown environment, based on the A* and Dijistra algorithm to complete the safety inspection in the known environment, collect, record and save the video during the inspection process,and based on YOLOv3, the pedestrian target in the video for detection, the Deep_Sort target tracking algorithm establishes tracking of the dectected target, and creates a learning sample data set 2019.