Modified Background Subtraction Research Articles

An Efficient Adjacent Frame Fusion Mechanism for Airborne Visual Object Detection

With the continuous advancement of drone technology, drones are demonstrating a trend toward autonomy and clustering. The detection of airborne objects from the perspective of drones is critical for addressing threats posed by aerial targets and ensuring the safety of drones in the flight process. Despite the rapid advancements in general object detection technology in recent years, the task of object detection from the unique perspective of drones remains a formidable challenge. In order to tackle this issue, our research presents a novel and efficient mechanism for adjacent frame fusion to enhance the performance of visual object detection in airborne scenarios. The proposed mechanism primarily consists of two modules: a feature alignment fusion module and a background subtraction module. The feature alignment fusion module aims to fuse features from aligned adjacent frames and key frames based on their similarity weights. The background subtraction module is designed to compute the difference between the foreground features extracted from the key frame and the background features obtained from the adjacent frames. This process enables a more effective enhancement of the target features. Given that this method can significantly enhance performance without a substantial increase in parameters and computational complexity, by effectively leveraging the feature information from adjacent frames, we refer to it as an efficient adjacent frame fusion mechanism. Experiments conducted on two challenging datasets demonstrate that the proposed method achieves superior performance compared to existing algorithms.

Vehicles detection and counting based on internet of things technology and video processing techniques

Recent studies have proven that vehicle tracking and detection play an important role in traffic density monitoring. Traffic overcrowding can be effectively controlled if the number of vehicles expected to pass through a congested intersection can be predicted ahead of time. To overcome such impact of traffic congestion the proposed system presents a framework, using motion detection algorithms and “ThingSpeak” internet of things (IoT) platform which is used in to calculate traffic density, the proposed system capturing video with wireless internet protocol (IP) cameras and broadcasting it to the server where motion detection algorithms as background subtraction are used to obtain a quick overview of traffic density, To save cost and improve the solution, the suggested system utilizes image processing techniques as well as the IoT analytic platform “ThingSpeak” to monitor traffic density. Finally, the suggested method is used to manage traffic flow and avoid traffic crowded. The results of the studies show that the integration of IoT-based technologies with a modified background subtraction technique is effective. This method might be enhanced further to detect vehicles that break traffic laws. We may also improve this system by detecting the presence of emergency vehicles (including an ambulance or fire truck) and granting priority to those cars.

Using modified background subtraction for detecting vehicles in videos

Using modified background subtraction for detecting vehicles in videos

Detection and tracking of moving object using modified background subtraction and Kalman filter

Moving object detection and tracking (MODT) is the major challenging issue in computer vision, which plays a vital role in many applications like robotics, surveillance, navigation systems, militaries, environmental monitoring etc. There are several existing techniques, which has been used to detect and track the moving object in Surveillance system. Therefore it is necessary to develop new algorithm or modified algorithm which is robust to work in both day and night time. In this paper, modified BGS technique is proposed. The video is first converted to number of frames, then these frame are applied to modified background subtraction technique with adaptive threshold which gives detected object. Kalman filter technique is used for tracking the detected object. The experimental results shows this proposed method can efficiently and correctly detect and track the moving objects with less processing time which is compared with existing techniques.

An Effective Method of Weld Defect Detection and Classification Based on Machine Vision

In order to effectively identify and classify weld defects of thin-walled metal canisters, a weld defect detection and classification algorithm based on machine vision is proposed in this paper. With the weld defects categorized, a modified background subtraction method based on Gaussian mixture models, is proposed to extract the feature areas of the weld defects. Then, we design an algorithm for weld detection and classification according to the extracted features. Next, by using the weld images sampled by the constructed weld defect detection system on a real-world production line, the parameters of the weld defect classifiers are determined empirically. Experimental results show that the proposed methods can identify and classify the weld defects with more than 95% accuracy rate. Moreover, the weld detection results obtained in the actual production line show that the detection and classification accuracy can reach more than 99%, which means that the system enhanced with the proposed method can meet the requirements for the best real-time and continuous weld defect detection systems available nowadays.

Anfis Classifier Based Moving Object Detection and Segmentation in Indoor and Outdoor Environments

Background: Moving object detection in dynamic environment video is more complex than the static environment videos. In this paper, moving objects in video sequences are detected and segmented using feature extraction based Adaptive Neuro-Fuzzy Inference System (ANFIS) classifier approach. The proposed moving object detection methodology is tested on different video sequences in both indoor and outdoor environments. Methods: This proposed methodology consists of background subtraction and classification modules. The absolute difference image is constructed in background subtraction module. The features are extracted from this difference image and these extracted features are trained and classified using ANFIS classification module. Results: The proposed moving object detection methodology is analyzed in terms of Accuracy, Recall, Average Accuracy, Precision and F-measure. The proposed moving object segmentation methodology is executed on different Central Processing Unit (CPU) processor as 1.8 GHz and 2.4 GHz for evaluating the performance during moving object segmentation. At present, some moving object detection systems used 1.8 GHz CPU processor. Recently, many systems for moving object detection are using 2.4 GHz CPU processor. Hence, CPU processors 1.8 GHz and 2.4 GHz are used in this paper for detecting the moving objects in video sequences. Table 1 shows the performance evaluation of proposed moving object detection on CPU processor 1.8 GHz (100 sequence). Table 2 shows the performance evaluation of proposed moving object detection on CPU processor 2.8 GHz (100 sequence). The average moving object detection time on CPU processor 1.8 GHz for fountain sequence is 62.5 seconds, for airport sequence is 64.7 seconds, for meeting room sequence is 71.6 seconds and for Lobby sequence is 73.5 seconds, respectively, as depicted in Table 3. The average elapsed time for moving object detection on 100 sequences is 68.07 seconds. The average moving object detection time on CPU processor 2.4 GHz for fountain sequence is 56.5 seconds, for airport sequence is 54.7 seconds, for meeting room sequence is 65.8 seconds and for Lobby sequence is 67.5 seconds, respectively, as depicted in Table 4. The average elapsed time for moving object detection on 100 sequences is 61.12 seconds. It is very clear from Table 3 and Table 4; the moving object detection time is reduced when the frequency of the CPU processor increases. Conclusion: In this paper, moving object is detected and segmented using ANFIS classifier. The proposed method initially segments the background image and then features are extracted from the threshold image. These features are trained and classified using ANFIS classification method. The proposed moving object detection method is tested on different video sequences which are obtained from different indoor and outdoor environments. The performance of the proposed moving object detection and segmentation methodology is analyzed in terms of Accuracy, Recall, Average Accuracy, Precision and F-measure.

Non-Line-of-Sight Optical Camera Communication in a Heterogeneous Reflective Background

In the last few years, optical camera communication (OCC) has benefited from built-in CMOS image sensors in smart devices. OCC is now considered a candidate for incorporation into IEEE 802.15.7r1. A major study on OCC was conducted for line-of-sight (LOS) environments. To enable users to receive OCC signals reflected from any surface easily, we considered non-LOS (NLOS) OCC in a heterogeneous reflective background in this study. A modified background subtraction method was proposed to identify the transmitted modulated signal that was recorded in an image. Then, the majority vote principle was used to restore the damaged pixels that were primarily caused by nonuniform reflection from the heterogeneous reflective background. The experimental results obtained after using a self-built NLOS OCC system demonstrated a significant improvement in the bit error rate when the proposed OCC signal recovery method was used for different distances between a smartphone and a heterogeneous reflective background.

Elephant detection using boundary sense deep learning (BSDL) architecture

ABSTRACT Elephant entry in human settlements causes a major threat to human–elephant conflict (HEC), but the rapid detection thereof while crossing boundary remains a difficult role, due to the lack of necessary boundary sensing infrastructure. A combination of wireless sensor networks and recent advances in deep learning has paved an effective way for elephant detection. This paper presents and evaluates a Boundary Sense Deep Learning architecture (BSDL) for automated detection of elephants, integrating, elephant image feature learning and elephant classification. A novelty in this approach enhances the deep learning Boundary Sense (BS) architecture to include a background subtraction module which highlights the visual region important for elephant recognition. The objective of the proposed BSDL architecture is validated by achieving better accuracy when compared with the raw image, time efficiency, simplifying and boosting up of Deep Convolution Neural Network (DCNN) by introducing preliminaries; elephant recognition with Multi view dataset dramatically improves the state of art and fine tuning using back-propagation. The dataset includes 1,28,069 images of various postures of elephants collected under different circumstances from different regions (Hosur, Bannerghatta, Shimoga (Sakrebailu) and Dubare). The trained model is designed for identifying only one target species, the elephants. The conclusions drawn from our work prove that the achievement percentage is 97% accuracy in the BSDL architecture, when compared with the existing deep learning algorithms using raw image.

Modified Background Subtraction Statistic Models for Improvement Detection and Counting of Active Spermatozoa Motility

An important early stage in the research of sperm analysis is the phase of sperm detection or separating sperm objects from images/video obtained from observations on semen. The success rate in separating sperm objects from semen fluids has an important role for further analysis of sperm objects. Algorithm or Background subtraction method is a process that can be used to separate moving objects (foreground) and background on sperm video data that tend to uni-modal. In this research, some of the subproject model statistics of substrata model are Gaussian single, Gaussian Mixture Model (GMM), Kernel Density Estimation and compared with some basic subtraction model background algorithm in detecting and counting the number of active spermatozoa. From the results of the tests, the Grimson GMM method has an f-measure value of 0.8265 and succeeded in extracting the sperm form near its original form compared to other methods

Using modified background subtraction for detecting vehicles in Videos

In this paper, a comparison study has been introduced between the traditional foreground detector (background subtraction technique) and a modified background subtraction-based detector (empty frame subtraction technique). A case study for such analysis has been introduced for estimating the average vehicular speed and the level of crowdedness in three test traffic videos with five different indices; frame rate, resolution, number of frames, duration, and extension). The proposed modification in the background subtraction detector strategy aims to reduce vehicle detection processing time which increase vehicle tracking efficacy. In addition, video degradations (salt and pepper noise, Gaussian noise, and speckle noise) have been applied in both traditional and modified background subtraction. Results have reflected an obvious decrease in the processing time by almost 40% than the traditional background detector.

Improved pedestrian detection using motion segmentation and silhouette orientation

Pedestrians are the most interesting as well as vulnerable entity from both safety and security perspective in the field of video surveillance. In this article, we present a framework to detect pedestrians across a stationary camera view. Our propositions thrust upon developing a motion segmentation module and a feature extraction module for human localization. In the first stage, a background subtraction method is proposed to collect the initial set of moving objects in the processed frame. A shape descriptor is then presented to encode the pattern of human body in terms of silhouette orientation histogram. Moreover, the principle of Golden ratio is employed to formulate a part-based detector to alleviate the problem with occlusion. Both the above modules are first validated separately, and then as a unified unit, using various statistical measures. The proposed background subtraction module is simulated on twenty video clips taken from three benchmark datasets. The efficacy of our shape descriptor is validated on various image-windows taken from three publicly available datasets. The unified framework including both the modules are tested on two standard surveillance datasets. All the experimental results are uploaded at: https://sites.google.com/site/PedestriansInMotion .

BBBCO and fuzzy entropy based modified background subtraction algorithm for object detection in videos

Background subtraction (BS) is one of the most commonly used methods for detecting moving objects in videos. In this task, moving objectpixels are extracted by subtracting the current frame from a background frame. The obtained difference is compared against a threshold value to classify pixels as belonging to the foreground or background regions. The threshold plays a crucial role in this categorization and can impact the accuracy and preciseness of the object boundaries obtained by the BS algorithm. This paper proposes an approach for enhancing and optimizing the performance of the standard BS algorithm. This approach uses the concept of fuzzy 2-partition entropy and Big Bang–Big Crunch Optimization (BBBCO). BBBCO is a recently proposed evolutionary optimization approach for providing solutions to problems operating on multiple variables within prescribed constraints. BBBCO enhances the standard BS algorithm by framing the problem of parameter detection for BS as an optimization problem, which is solved using the concept of fuzzy 2-partition entropy. The proposed method is evaluated using videos from benchmark datasets and a number of statistical metrics. The method is also compared with standard BS and another recently proposed method. The results show the promise of the proposed method.

A Computationally Economic Novel Approach for Real-Time Moving Multi-vehicle Detection and Tracking toward Efficient Traffic Surveillance

Efficient and precise real-time video-based vehicles supervision systems for traffic surveillance are ever demanding. Developing an accurate detection and tracking algorithm by incorporating the effects of vehicles shadow or visual obstacles/occlusion including road signs, trees, or other vehicles is a challenging task. We propose a new approach for efficient detection and tracking of moving vehicles in the presence of shadow and partial occlusion under complex road scenes by combining the modified background subtraction and innovative adaptive search window methods. The use of variable-size window (Bounding Box) reduces the searching space proportional to the allocated time of the moving vehicles. First, a background model is constructed and then subtracted from subsequent traffic frames to distinguish the foreground objects (moving vehicles regions) and stable background regions. Finally, the shadow detection and occlusion handling between the moving vehicles are implemented to leverage the surveillance performance. The performance of the proposed method is benchmarked using baseline highway category of changedetection.net standard dataset having 1231 frames. Performance evaluation is carried out in terms of time complexity analysis, detection accuracy, and processing time. The experimental results demonstrate a radical reduction in computational time in terms of searched space. A superior detection accuracy (more than 99%), and processing time \(\sim \)29.6 ms (per frames in milliseconds) together with the time complexity degree of O(\(n^{2}\)), is achieved. Excellent features of these results suggest that the proposed approach is prospective for computer vision application and high-way traffic surveillance.

Hardware Design of Moving Object Detection on Reconfigurable System

Moving object detection including background subtraction and morphological processing is a critical research topic for video surveillance because of its high computational loading and power consumption. This paper proposes a hardware design to accelerate the computation of background subtraction with low power consumption. A real-time background subtraction method is designed with a frame-buffer scheme and function partition to improve throughput, and implemented using Verilog HDL on FPGA. The design parallelizes the computations of background update and subtraction with a seven-stage pipeline. A stripe-based morphological processing and accounting for the completion of detected objects is devised. Simulation results for videos of VGA resolutions on a low-end FPGA device show 368 fps throughput for only the real-time background subtraction module, and 51 fps for the whole system, including off-chip memory access. Real-time efficiency with low power consumption and low resource utilization is thus demonstrated.

An Automated Attendance System: A Technique using Background Subtraction and Color Image Processing of RGB Planes

Background/Objectives: The objective of this paper is to introduce an automated attendance marking system for LPU using CCTV cameras. It detects the students sitting in the class and takes the attendance automatically without any human assistance and count total number of students present in the class also. Methods/Statistical Analysis: An enhanced detection algorithm is proposed that is having two phases: color image processing of RGB planes and modified background subtraction of image which executed simultaneously for object detection. Before detection process, noise is removed from images by using log function which uniformly distributes the brightness over the image. To count the number of detected objects, a counter is used which calculates total number of objects present in an image. The Automatic ROI detection algorithm is also proposed which detects region of interest automatically on the basis of repeated regular pattern of object. Findings: The system able to count and mark attendance of students sitting in class room. Implementation shows that our proposed approach has 89.5% accuracy with less false positive and false negative rate as compared to other traditional approaches. The Automatic ROI detection algorithm also detects the region of interest with considerably high accuracy. Application/Improvements: The proposed system will easy to deploy in various schools, colleges, universities and other educational institutes with minimal setup cost where CCTV cameras are already installed in class rooms.

Monitoring Debris Flows Using Spatial Filtering and Entropy Determination Approaches

We developed an automatic debris flow warning system in this study. The system uses a fixed video camera mounted over mountainous streams with a high risk for debris flows. The focus of this study is to develop an automatic algorithm for detecting debris flows with a low computational effort which can facilitate real-time implementation. The algorithm is based on a moving object detection technique to detect debris flow by comparing among video frames. Background subtraction is the kernel of the algorithm to reduce the computational effort, but non-rigid properties and color similarity of the object and the background color introduces some difficulties. Therefore, we used several spatial filtering approaches to increase the performance of the background subtraction. To increase the accuracy entropy is used with histogram analysis to identify whether a debris flow occurred. The modified background subtraction approach using spatial filtering and entropy determination is adopted to overcome the error in moving detection caused by non-rigid and similarities in color properties. The results of this study show that the approach described here can improve performance and also reduce the computational effort.

Image Segmentation in Video Sequences Using Modified Background Subtraction

In computer vision, Background is a technique for finding moving objects in a video sequences for example vehicle driving on a freeway. For to detect non stationary (dynamic) objects, it is necessary to subtracting current image from a time-averaged background image. There are various background subtraction algorithms for detecting moving vehicles or any moving object(s) like pedestrians in urban traffic video sequences. A crude approximation to the task of classifying each pixel on the frame of current image, locate slow-moving objects or in poor image qualities of videos and distinguish shadows from moving objects by using modified background subtraction method. While classifying each pixel on the frame of the current image, it is to be detect the moving object at foreground and background conditional environment that we can classify each pixel using a model of how that pixel looks when it is part of video frame classes. A mixture of Gaussians classification model for each pixel using an unsupervised technique is an efficient, incremental version of Expectation Maximization (EM) is used for the purpose. Unlike standard image-averaging approach, this method automatically updates the mixture component for each video frame class according to likelihood of membership; hence slowmoving objects and poor image quality of videos are also being handled perfectly. Our approach identifies and eliminates shadows much more effectively than other techniques like thresholding.

A smart camera application: DSP-based people detection and tracking

As an application of smart cameras, this paper presents an algorithm and its real-time implementation for detecting and tracking multiple people moving in a surveillance scene. The algo- rithm is devised to work for low resolution image frames and any camera angle. First an adaptive background subtraction module is used to obtain potential moving target areas. Second a human body detection module is activated to locate the heads. Third a multitarget tracking module is deployed to resolve ambiguities associated with crosspaths and merged targets. The algorithm is implemented on the Texas Instruments TMS320C6201 fixed-point DSP processor for real-time processing of video signals. This real-time capability allows field deployment for the purpose of collecting surveillance data such as the number of people and their crossing frequency in a certain area. © 2000 SPIE and IS&T. (S1017-9909(00)01802-X)