Human Motion Detection based on Background Subtraction Techniques - Coursework Example

Human Motion Detection based on Background Subtraction techniques In most computer vision applications, detection and identification of moving objects becomes the fundamental requisite. Background subtraction also known as foreground detection is a technology used in the fields of computer vision and image processing. It is a common approach for detecting moving objects in videos from static cameras by subtracting the current image from the background image or background module. Extensive research has been done in this field to accurately extract the motion region of interest in order to use it for further processing such as object recognition. This paper provides a review of the human motion detection methods focusing on background subtraction technique. keywords: motion detection method, background subtraction, moving object detection. 1- Introduction Background subtraction also known as foreground detection is a common approach for detecting moving objects in videos from static cameras. It widely used for traffic monitoring, human action recognition, human computer interaction, object tracking and computer vision application. The basic operation needs is the separation of the moving objects called the foreground from the static information called the background the process mainly used is the background subtraction approach [1]. The background image must be a representation of the scene with no moving objects and must be kept regularly updated so as to adapt to the varying luminance conditions and geometry settings. More difficult models have extended the concept of “background subtraction” beyond its literal meaning. The background subtraction method is the common method of motion detection. It is a technology that uses the difference of the current image and the background image to detect the motion region, and it is generally able to provide data included object information . the simplest way is to apply the difference between the two frames as follow : | I(X,Y,t) - I(X,Y,t-1) | > Th (1) where I(X,Y,t-1) is the background frame referred as previous frame subtracted from the current frame I(X,Y,t), then apply the threshold (Th) to the absolute difference to get the foreground mask.where the key parameter in the thresholding process is the choice of the threshold value, and it can be chosen manually or automatically. the rationale in the approach is that of detecting the moving objects or foreground from the compression between the input video frame and the background frame. in this paper however, any motion detection system based on background subtraction has been taken into account a number of considerations such as : background subtraction should segment objects of interest when they first appear in a scene. An appropriate pixel-level stationarity criterion should be defined. pixel that satisfy this criterion are declared background and ignored the background model must adapt to both sudden and gradual changes in the background and light conditions in the scene. background models should take into account changes at differing spatial scales. Shadow regions that are projected by foreground objects and are detected as moving objects. Multiple objects moving in the scene both for long and short periods. the main objective of this paper is to provide an algorithm that can detect human motion at certain distance for object tracking application. it carry out various tasks such as motion detection, background modeling and subtraction, shadow detection and removal. 2- Literature review 2-1 Motion detection Method The motion detection methods are classified according to the method of finding moving object [2]. Different motion detection methods are described as follows: A- temporal difference The simplest form of background subtraction is the Frame difference. It refers to the subtraction of the current frame from the previous frame and incase the pixel values for a given pixel exceeds the threshold (Th) value, the particular pixel is considered to be a part of the foreground [3]. Temporal difference has high adaptability especially for a wide array of dynamic environments, however it is generally difficult to derive an accurate outline of a moving object and the movement detection obtained becomes faulty. B- optical flow The optical flow method is employed to calculate the image optical flow field and conduct clustering processing in accordance to the optical flow distribution characteristics of image [4]. The optical flow methodology can help in obtaining clearer and accurate information about the detection and movement of objects against the background. However owing to the method’s sensitivity to noise and poor anti-noise performance, it is not suitable for studies demanding real-time detection [5]. C- background subtraction Background subtraction is a widely used technique for segmentation of motion in static images. It detects moving regions by subtracting the required current image pixel-by-pixel from a reference background image which is created by averaging several images during an initial period. The primary idea of background subtraction is detect a moving object by first initializing a background and then subtracting the current requisite frame containing the moving object to be detected. The background subtraction method is simple and follows a easy protocol but even though it accurately helps obtain characteristics of target data, it is sensitive to minute changes in the external environment and therefore it can only be used in those conditions where the background is known or maybe predicted [4]. 3- Motion detection system design The moving object in the frame sequences is detected using background subtraction. For motion detection, two images preferably of the same size are taken from video. In that one image is initialized as the background image in which the moving object is not present and the second image is the current image. And each image has two models one is the foreground and background model. The Fig.1 shows the system block diagram. figure 1: System Block Diagram The foreground model is the model in which the moving object is present and background model is the model in which the moving object is not present. The first process for motion detection is image initialisation. Image initialization is process that initializes the background image. For example, in the video the number of the frames with respect to the time, out of these frame one is initialized as the background image by taking some assumption. Hence initialization of background is essential preprocessing operation for motion detection. And the preprocessing is done on each frame and the preprocessing is done by the mean or median filter for reducing the noise from the image. After the preprocessing the frames are given to the background subtraction algorithm. That subtracted image is then segmented using Thresholding. 5- background subtraction method ………... 5.1 Background Image Initialization There are many methods to obtain the initial background image. For instance, with the first frame as the background directly or, the average pixel brightness of the first few frames as the background or using a background image sequences without the hope of moving objects to approximate the background model parameters and so on. From these methods average method is commonly used for background Image initialization [3], but there are many shadow problems will occur which can be removed by median method, Expression is as follows: Binit (X,Y) = median Fk(X,Y) k =1,2 ….n (2) Where Binit is the initial background, n is the total number of frames selected. 5.2 Background Update( model) For the background model can better adapt to light changes, the background needs to be updated in real time, so as to accurately extract the moving object. In this paper, the update algorithm is as follows: In detection of the moving object, the pixels judged as belonging to the moving object maintain the original background gray values, not be updated. For the pixels which are judged to be the background, we update the background model according to following rules: Bk+1(x,y)= beta Bk(x,y)+(1-beta)Fk(x,y) (3) where beta belongs to (0,1) is update coefficient , in this paper beta=0.004, Fk(x,y) is the pixel grey value in the current frame. Bk(x,y) and Bk+1(x,y) are respectively the background value of the current frame and Next frame.As the camera is fixed, the background model can remain relatively stable in the long period of time. Using this method can effectively avoid the unexpected phenomenon of the background, such as the sudden appearance of something in the background which is not included in the original background. Moreover by the update of pixel gray value of the background, the impact brought by light, weather and other changes in the external environment can be effectively adapted. 5.3 moving object extraction/detection Background subtraction is a popular technique to fragment out the interested objects in a frame. This technique involves subtracting an image that contains the object, with the previous background image that has no foreground objects of interest. The area of the image plane where there is a significant difference within these images indicates the pixel location of the moving objects[5]. These objects, which are represented by groups of pixel, are then separated from the background image by using threshold technique. After the background image B(X,Y) is obtained, subtract the background image B(X,Y) from the current frame Fk(X,Y).If the pixel difference is greater than the set threshold Th , then determines that the pixels appear in the moving object, otherwise, as the background pixels. The moving object can be detected after threshold operation. Expression is as follows: Dk(X,Y)= 1 if | Fk (X,Y)-Bk(X,Y) | > Th ((((( or add dynamic threshold)))) 0 Otherwise Human body detection is to identify the corresponding part of human from the moving region. But the extracted moving region may correspond to different moving objects, such as vehicles and other such birds, floating clouds, swaying tree and other moving objects [8]. Hence we use the shape features of motion regions to further determine whether the moving object is a human being or not. Judging criteria are as follows: 1. The object area is larger than the set threshold 2. The aspect ratio of the object region should conform to the set ratio. If these two conditions are met, the moving object is the moving human body or not a human body is conform. 6.2 Noise Removal ( reprocessing) Since the difficulty of the background, the discrepancy image obtained contains the motion region as well as large number of noise. These noises might be included in the image due to environmental factors, illumination changes, during transmission of video from the camera to the further processing. Therefore, noise needs to be removed. median filter can be adopted for filtering noise. As we know, motion region just not includes human being, but also it may include moving cars, flying birds, flowing clouds and swaying trees and other non body parts. Morphological methods could be used for further processing. Corrosion operation is taken to effectively filter out non-human activity areas and by using the development operation they can filter out most of the non-body motion regions while preserving the shape of human motion without injury. After expansion and corrosion operations, some inaccessible spots of the image and some intrusion of small pieces are eliminated, and we get more accurate human motion region. 6.3 Extraction of Moving Human Body( shape analysis) Some accurate edge regions will be got after median filtering, corrosion and expansion operations, but the region belongs to the moving human body could not be determined. Through inspection, we can find out that when moving object appears, shadow will appear in some regions of the scène.Accurate mining of the moving object affected by the presence of shadow. By analyzing the characteristics of motion detection, we combine the projection operator with the previous methods. Based on the results of the methods above, height of the motion region will get detected by adopting the method of combining horizontal with vertical projection. This can eliminate the impact of the shadow to a certain level. Then we analyze the vertical projection value and set the threshold value to remove the pseudo-local maximum value and the pseudo-local minimum value of the vertical projection to determine the number and width of the body in the motion region, we will get the moving human body with precise edge. 6.4 Behavior Understanding After successfully detecting the moving humans from one frame to another in an image sequence, the problem of understanding human behaviors from image sequences follows naturally. Behavior understanding involves action identification and description. Human behavior understanding can guide the development of many human motion analysis systems. Behavior understanding is to analyze and recognize human motion patterns, and to produce high-level description of actions and interactions .The behavior understanding will be the most important area of future research in human motion analysis. 5-conclusion Background subtraction is a common method used for human motion detection. This review paper described the basic approach for background subtraction method. Median filter and morphological operations can be used to eliminate the noise during the process. However, more robust methodologies to eliminate environment noise needs to be the focus of current research. In addition to this proper background modeling needs to be done to make the process efficient in real-time. 6-references [1]traditional approaches in background modeling for static cameras, Thierry Bouwmans 2015. [2]Detection Of Moving Object Based On Background Subtraction, Mr. Mahesh C. Pawaskar1, Mr. N. S.Narkhede2 and Mr. Saurabh S. Athalye1, 2014 [3] M.Kalpana Chowdary , S.Suparshya Babu, S.Susrutha Babu, Dr.Habibulla Khan “FPGA Implementation of Moving Object Detection in Frames by Using Background Subtraction Algorithm” International conference on Communication and Signal Processing, April 3-5, 2013, India ©2013 IEEE pp 1032-1036. [4] K.Kinoshita, M.Enokidani, M. Izumida and K.Murakami, "Tracking of a Moving Object Using One-Dimensional Optical Flow with a Rotating Observer," Control, Automation, Robotics and Vision, 2006. ICARCV 06. 9th International Conference on 5-8 Dec. 2006 pp 1 – 6 [5] Human Motion Detection using Background Subtraction Algorithm R.S.Rakibe* Prof.B.D.Patil, 2014 [6]Background Subtraction Algorithm Based Human ,Rupali S.Rakibe*, Bharati D.Patil** ,2013 [7] B. S. M. Madhavi and M. V. G. Rao, “a Fast and Reliable Motion Human Detection and Tracking Based on Background Subtraction,” vol. 1, no. 1, pp. 29–35, 2012. Read More