Early fire detection method in video for vessels

New generation vessels are equipped with fire detecting sensors; however, fire may not immediately be detected if it is far away from the sensors. The fire process therefore cannot be recorded. A video-based fire alarm system is developed to overcome the drawbacks of traditional fire detection equipment. This paper presents a video-based flame and smoke detection method for vessels. For flame detection, the dominant flame color lookup table (DFCLT) is created by using the fuzzy c-means clustering algorithm. The changed video frames are automatically selected and the changed regions deduced from these frames. An elementary, medium, or emergency flame alarm is then triggered by comparing the pixels of changed regions with the DFCLT. The changed video frames are automatically selected for smoke detection. The changed regions are deduced from these frames. If the shape of the changed region conforms to the characteristic which the top area is wider than the bottom area, a dangerous smoke alarm is sounded. The experimental results show that the proposed fire detection approach can detect dangerous flames and smoke, effectively and efficiently.     

基于视频图像的火灾自动检测

首先分析了火焰的基本特性,给出了火焰图像的特征描述,然后利用火焰图像序列的边缘不稳定和相似性等可识别特征以及面积大小和颜色等信息,实现了对视频序列图像中火焰的自动检测。实验证明,所实现的系统检测效果好、误判率低。     

视频火灾检测的设计与实现

在分析火灾图像特性的基础上,运用数字图像处理技术和模式识别技术,提出了火灾识别的思想.给出了图像处理和识别的算法,该算法采用二维最大熵自动阈值法对火灾图像进行分割处理,分割后再提取可疑区域;对可疑区域的火焰进行识别,给出火焰存在的可能性;根据火灾火焰蔓延时的面积、相似度的变化来识别、判断火灾的发生.实验证明,与传统的检测方法相比,大大地提高火灾预报的准确率.     

A fast self-similarity matrix-based method for shrew DDoS attack detection

ABSTRACT

Shrew DDoS attack mainly targets the TCP’s retransmission timeout (RTO) mechanism that handles severe cases of congestion and packet losses. This attack is very hard to detect due to its stealthy nature and low-rate in volume which if remained undetected can affect the legitimate TCP flows. In this paper, we propose a fast shrew DDoS attack detection method based on self-similarity matrix (SSM) that measures the self-similarity of network traffic across multiple time scales over a subset of relevant features. The method can detect any presence of shrew attack in-line with the incoming traffic samples and thus identify the attack flows. We experimented our method over real-life low-rate datasets for multiple scenarios and the results demonstrate its efficiency both in terms of detection accuracy and speed.     

Fire flame detection in video sequences using multi-stage pattern recognition techniques

In this paper, we propose an effective technique that is used to automatically detect fire in video images. The proposed algorithm is composed of four stages: (1) an adaptive Gaussian mixture model to detect moving regions, (2) a fuzzy c-means (FCM) algorithm to segment the candidate fire regions from these moving regions based on the color of fire, (3) special parameters extracted based on the tempo-spatial characteristics of fire regions, and (4) a support vector machine (SVM) algorithm using these special parameters to distinguish between fire and non-fire. Experimental results indicate that the proposed method outperforms other fire detection algorithms, providing high reliability and a low false alarm rate.     

井下视频行人检测方法

针对现有基于深度学习的行人检测方法存在计算量较大、检测效率严重依赖硬件性能等问题,对基于SSD网络的行人检测方法进行改进,设计了一种基于DenseNet网络的轻量级卷积神经网络作为SSD网络的基础网络,以满足井下视频行人实时检测需求,并设计了基于ResNet网络的辅助网络,以增强特征表征能力,提高行人检测准确性。将基于改进SSD网络的井下视频行人检测方法部署在嵌入式平台Jetson TX2上进行实验,结果表明该方法对井下视频中行人的检测准确率为87.9%,针对井下行人低密度场景的检测准确率近100%,且运算速度达48帧/s,约为基于SSD网络的行人检测方法的4.4倍,满足井下行人实时检测需求。     

基于视频的烟焰检测方法研究

烟焰作为火灾最为明显的标志,对于其进行实时检测,在火灾监控方面有着重要的意义.传统的离子式、光电式、吸烟式等烟雾探测器受到空间和时间的限制,不能进行实时全面的探测[1].本文提出了一种基于视频的烟焰检测算法,采用帧差法提取出差值图像,经过图像处理,获得了理想的烟焰轮廓,根据轮廓大小来确定被监测区域是否有火灾的发生.实验证明,该方法实时准确、灵敏度高、抗干扰能力强、适用范围广,具有广阔的应用前景.     

基于视频的改进的火焰目标分割方法

本文提出了一种改进的基于RGB与HSI颜色模型的火焰目标分割方法,能完整地提取出火焰目标。这种方法是在实时视频流中,采用对抽取出的两帧图像进行帧间差;然后根据RGB与HSI火焰颜色模型进行筛选,获取火焰目标边缘部分像素作为火焰种子;进行以火焰种子遍历八邻域区域生长的方式,提取出完整的火焰。相比以往的火焰分割方法,改进后的方法能更完整地对火焰目标实现分割,也更满足实时视频监控的要求。实验结果证实,提出的改进方法能完整地分割出火焰目标。     

Real-time video fire smoke detection by utilizing spatial-temporal ConvNet features

Fire is one of the most dangerous disasters threatening human life and property globally. In order to reduce fire losses, researches on video analysis for early smoke detection have become particularly significant. However, it is still a challenging task to extract stable features for smoke recognition, largely due to its variations in color, shapes and texture. Classical convolutional neural networks can automatically learn feature representations of appearance from a single frame but fail to capture motion information between frames. For addressing this issue, in this paper, we propose a spatial-temporal based convolutional neural network for video smoke detection, and for real-time detection, propose an enhanced architecture, which utilizes a multitask learning strategy to jointly recognize smoke and estimate optical flow, capturing intra-frame appearance features and inter-frame motion features simultaneously. The effectiveness and efficiency of our proposed method is validated by experiments carried out on our self-created dataset, which achieves 97.0% detection rate and 3.5% false alarm rate with processing time of 5ms per frame, obviously outperforming existing methods.     

基于视频的火焰检测方法

在分析视频监控的火焰检测技术的基础上,针对传统的帧间差分法,提出一种结合分块处理和帧间差分的分块帧组差分的运动检测方法,提高相邻像素间的关联性,降低噪声干扰的影响,通过色调、色饱和度和亮度来分析视频中火焰像素的静态颜色特征,降低亮度对颜色检测的影响.充分利用燃烧火焰的动态频谱特征和空间的能量信息,对不同时刻火焰燃烧的图像时域和空域进行小波分析;分析后的疑似火焰区域利用连通邻域像素的信息对零散的非火焰像素点进行滤除.实验结果表明,该检测方法更具目的性,且颜色检测更具可靠性,降低了闪烁光和类似火焰物体等造成的影响.     

A new PSO-based approach to fire flame detection using K-Medoids clustering

Automated computer vision-based fire detection has gained popularity in recent years, as every fire detection needs to be fast and accurate. In this paper, a new fire detection method using image processing techniques is proposed. We explore how to create a fire flame-based colour space via a linear multiplication of a conversion matrix and colour features of a sample image. We show how the matrix multiplication can result in a differentiating colour space, in which the fire part is highlighted and the non-fire part is dimmed. Particle Swarm Optimization (PSO) and sample pixels from an image are used to obtain the weights of the colour-differentiating conversion matrix, and K-medoids provides a fitness metric for the PSO procedure. The obtained conversion matrix can be used for fire detection on different fire images without performing the PSO procedure. This allows a fast and easy implementable fire detection system. The empirical results indicate that the proposed method provides both qualitatively and quantitatively better results when compared to some of the conventional and state-of-the-art algorithms.     

An integrated fire detection and suppression system based on widely available video surveillance

An integrated system based on video surveillance is presented for automatic fire detection and suppression. The system is composed of two modules, including fire detection and fire suppression. The fire detection module makes full use of traditional CCD cameras for fire recognition. Some spatio-temporal features, such as color and motion, are extracted to detect fire in real time by utilizing sequential image processing techniques. Once a fire is detected, the system will control the fire suppression module to extinguish the fire automatically. It mainly consists of control device, mobile device, and water gun. Experiments performed in a large space hall show that the integrated system can detect a fire about a few seconds after ignition and automatically suppress the fire quickly.     

一种火焰检测系统的信号处理方法

设计了一种用于热能设备的火焰检测系统处理信号的方法.系统以AVR单片机为核心芯片,信号放大处理电路采用多级放大器与多路ADC通道相结合,实现自适应增益调整,以提高输入动态范围.火焰信号强度用输入信号的统计平均值表示,火焰信号的频率用快速傅里叶变换(FFT)进行分析计算.在FFT运算中,利用定点运算和查表法简化计算过程,降低资源消耗,提高运算速度,在300 ms时间内完成了512个采样点的定点FFT运算,频率分辨率为1 Hz,保证火焰信号检测的实时性、精度和可靠性.     

A novel matrix-based method for face recognition

Two-dimensional principal component analysis (2DPCA) is one of the representative techniques for image representation and recognition. However, it fails in detecting the local variation of images, which characterizes the most important modes of variability of face images. Motivated by the fact that the local spatial geometric structure of images is effectual in learning the representative image space, we assign different weight to each training image and then present a novel method, namely local two-dimensional principal component analysis (L2DPCA), which explicitly considers the variations among nearby data. Finally, we describe an effective algorithm L2DPCA+2DPCA to further reduce dimensionality reduction. Extensive experimental results on two-face databases (Yale and AR) show the efficiency of the proposed method.     

A threshold-adaptive film mode detection method in video de-interlacing

With the popularization of the network video, format conversion and de-interlacing are more required in video displaying and transmission. Field merge is perfect to de-interlace the film mode video sequence in which repeated fields or even/odd fields from the same film frame are contained. In this paper, inter-field difference is compared with a threshold, and repetition pattern of the result is utilized to determine the film mode. The threshold is set variable by tracking the local minimum difference of previous and subsequent fields. Film modes with repeated fields are identified using frame difference pattern. The 2-2 film mode is decided if forward and backward field differences jaggedly swing in opposite direction. Other film modes with fields from the same film frame are recognized by the combination of frame difference and forward field difference patterns. In this way, correct field merge is guaranteed after knowing which adjacent fields are decomposed from the same film frame. Experimental results show that various film modes in television systems, animations, camcorders and particularly interlaced network video can be correctly recognized by proposed method.     

一种集成贝叶斯决策的视频烟雾检测新方法

研究将贝叶斯决策应用于自适应神经-模糊推理系统（ANFIS）的视频烟雾检测系统。提取视频烟雾特征,通过减法聚类和混合学习算法,确定并优化得到ANFIS实例,引入贝叶斯决策对ANFIS输出进行检测判别。仿真实验表明,ANFIS比其他烟雾检测算法具备更好的检测性能,而基于最小风险的贝叶斯决策可进一步提高检测率和降低虚警率,能更好地满足实际应用的需求。     

一种基于事件检测的视频取证方法*

目前计算机视频取证的一个重要目标是如何快速准确地在海量视频中定位犯罪事件发生的时刻和地点,最终形成视频证据。针对复杂背景条件下丢弃或拾起等事件的监控视频,提出一种基于光流特征和形状特征结合的事件检测方法。通过实验证明了该方法在视频事件分析取证中的有效性。     

视频图像中的视觉疲劳实时检测方法研究

提出一种新的适用于驾驶中视觉疲劳实时检测的人脸定位及眼睛状态分析算法。采用差分法快速找到视频图像中的目标运动区域,结合YCbCr色彩空间进行肤色分割定位人脸。对脸部区域进行灰度积分投影并结合Hough变换检测眼睑。对检测到的眼睑进行数据分析,得到眼睛开闭情况,结合眨眼分析,获得EOD值来判断驾驶员是否疲劳。实验结果显示该方法能在复杂背景下快速定位人脸,检测到眼睛睁开时的EOD值,满足视觉疲劳检测的实时需要。     

低空航拍视频中基于Vibe算法的船舶检测方法

针对低空航拍视频中的船舶检测研究普遍存在因背景变化及水纹扰动导致的识别效果不理想等问题,提出了一种基于改进型Vibe(Visual background extractor)算法的船舶检测方法。该方法在传统Vibe算法基础上融合了改进型的Canny算子,采取平滑滤波与自适应阈值分割策略提高了船舶轮廓提取的准确性。利用四方位结构元素的形态学机制,降低了水面波纹对目标识别的干扰影响。对于船舶检测中的内部空洞则选取了种子填充法进行填充,最后根据船舶外轮廓实现对船舶目标的框定。实验结果表明,所提出的改进型Vibe算法的船舶检测效果明显优于传统方法,验证了其可行性与实用性。