Showing papers on "Change detection published in 1990"

Scene change detection system and method

Abstract: A system detects scene changes in a sequence of video images by analyzing the sequence for abrupt frame-to-frame changes in certain image features. The system accepts the signal into a quantizer, which digitizes the image, and stores it into a frame buffer. An image processor, a component of the system, analyzes the digitized images, and determines certain features which a decision processor can use to detect a scene change.

Change detection in SAR imagery

Abstract: The author describes two separate SAR (synthetic aperture radar) image change detection schemes: (1) model based image segmentation (2) image object identification using neural networks. The neural network approach performs slightly better than the model based scheme in terms of both time and detection rates. >

Land use change detection based on multi-date imagery from different satellite sensor systems

Abstract: An empirical study is conducted to assess the accuracy of land use change detection using satellite image data acquired ten years apart by sensors with differing spatial resolutions. The primary goals of the investigation were to (1) compare standard change detection methods applied to image data of varying spatial resolution, (2) assess whether to transform the raster grid of the higher resolution image data to that of the lower resolution raster grid or vice versa in the registration process, (3) determine if Landsat/Thermatic Mapper or SPOT/High Resolution Visible multispectral data provide more accurate detection of land use changes when registered to historical Landsat/MSS data. It is concluded that image ratioing of multisensor, multidate satellite data produced higher change detection accuracies than did principal components analysis, and that it is useful as a land use change enhancement method.

Accurate detection of a drastic change between successive pictures

Abstract: On detecting a drastic change, such as a scene change, between current and previous ones of successive pictures which are represented by a digital video signal and each of which comprises a predetermined number of picture elements, the digital video signal is processed by using correlations between the picture elements within each of the successive pictures to produce a processed signal which represents processed results for each of the successive pictures. A statistical distribution of the processed results is calculated from the processed signal. The statistical distribution of the processed results in the current picture is collated with the statistical distribution of the processed results in the previous picture to be produced as a change detection signal representative of the drastic change with reference to a relationship between the statistical distribution of the processed results in the current picture and the statistical distribution of the processed results in the previous picture. The processed signal may be obtained by producing a difference signal between the digital video signal and a prediction signal predictive of the digital video signal by using the correlations. Alternatively, an orthogonal transform is carried out for the digital video signal to produce the processed signal.

Change point detection in a stochastic complexity framework

Abstract: The authors present a method, inspired by stochastic complexity theory, for solving the change point detection problem for ARMA (autoregressive moving average) systems which are assumed to have a slow unstructured nondecaying drift after the change has occurred. The central idea is to apply the minimum description length method in the form of predictive stochastic complexity, which gives a way for selecting the best model among a given set of models. Therefore the change point detection problem is reduced to a model selection problem. Simulations that show that the approach exhibits good detection capabilities are included. >

Change point detection method for moving image

Abstract: PURPOSE:To perform high-speed change point detection by determining the correlation coefficient of the feature amount by a frame unit with the feature amount of the last frame in the moving image composed of continuing plural images and deciding a point where the change rate of the correlation coefficient is beyond the prescribed value as the change point of a cut. CONSTITUTION:A picture input processing 21 fetches the moving image by one frame in a frame memory. Further, the address on the medium of the frame is read as a time code. An image preprocessing 22 performs the thinning of the image and the compression of a color picture element. A frame correlation coefficient calculation processing 23 performs the calculation of the frame corre lation coefficient with the last frame to store the result in an array. A change rate calculation processing 24 calculates the change rate of the correlation coefficient to be stored in a variable DDT. A judgement processing 25 decides the change of the cut when the change rate DDT is higher than the prescribed value. An image change point registration processing 26 prepares the index information of the moving image cut. A frame update processing 27 updates the frame.

Change detection in SAR imaginery

Abstract: The two major difficulties associated with SAR (synthetic aperture radar) image change detection are addressed. These are the removal of speckle noise and the registration of information between images. Due to the unpredictable nature of the aircraft track, the problem of image registration is severe in airborne SAR imagery. Autofocus techniques are used to measure residual aircraft motions, thus allowing the production of large geometrically accurate images. The problem of speckle reduction is approached in two ways. The first technique applies an intensity segmentation algorithm to each image. The regions generated by the segmentation are then compared by the change detection algorithm. An alternative approach is to use neural network methods to learn the speckle removal and region generation task. To reduce this problem to a manageable size a factorization method for the multilayer-perceptron has been invented. >

Real—time SAR change—detection using neural networks

Abstract: This paper describes the techniques evolved at RSRE for the production of undistorted, focused synthetic aperture radar (SAR) images, target detection using a neural network method and the automatic detection of changes between pairs of SAR images. All these processes are achievable in a single pipelined process operating on an input data rate in excess of 10 Mbytes/second.© (1990) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Adaptive multisensor change detection

Abstract: This paper describes an approach for intruder detection by a remote surveillance system. Emphasis is placed on the development of multispectral vision techniques for the extraction of information from a noisy and cluttered environment. This approach uses adaptive detection for operation under changing illumination and thermal environments. The system is initialized with an operator-guided segmentation to partition the scene into regions of similar noise characteristics and processing priorities. Images from a color TV and a FLIR are registered electronically and a common segmentation is used for both. Detection processing in corresponding regions of the two sensors' images are closely coupled. A system testbed is described and several processing sequences are presented which illustrate the approach.© (1990) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Diffusion approximation for a pseudo-likelihood test process with application to detection of change in a stochastic system

Abstract: We prove an in variance principle for a statistic related to an adaptive algorithm. The results obtained allow us to provide a rigorous mathematical justification ui the change detection and model validation procedure proposed by Basseville, Benveniste and Moustakides [2j. We also prove the optmiaiity, in a suitable statistical sense, of this procedure when applied to a model validation problem

Statistical analysis of a change detector based on image modeling of difference picture

Abstract: A statistical analysis of a change detector for motion detection based on image modeling of a difference picture is presented. The approach is founded upon the fact that complete silhouettes of moving objects can be segmented at each frame time using a change detector based on image modeling of the difference picture. The change detection problem can be treated as a signal detection problem in which very little is known about the signal to be detected. The exact modeling of the background event is necessary for a robust change detector which can adapt to changing environments. The governing statistics for change measures, which have a form of sum of squares of the difference picture, can be inferred from the assumed Gaussian distribution of the stationary background in the difference picture and are used for calculation of the adaptive decision threshold for the change detector. It is shown that the adaptive decision threshold derived from the change measure minimizes the total probability of error as measured in false alarms and missed detections. Experimental results that support the underlying statistical analysis are presented. >

Object Tracking and Background Estimation with a Moving Camera

Abstract: For navigation of vehicles or robots in an environment where other moving objects may be present, algorithms for detection and tracking of moving objects in image sequences are required that can be applied even if the camera itself is moving. A similar problem arises if the camera is mounted on a pole in an outside environment, e. g. for traffic monitoring, and is subject to mechanical disturbances causing jitter in the image sequence. In both cases a dynamical model for the egomotion of the camera can be used for motion estimation and motion compensation. The resulting procedure combines local image matching, Kalman filtering, autoregressive (AR) parameter estimation, background image estimation, and change detection for tracking the moving objects.

Very low rate coding of motion video using 3-D segmentation with two change-detection masks

Abstract: A new 3-D segmentation-based coding technique is proposed for transmitting the motion video with reasonablly acceptable quality even at a very low bit rate Only meaningful motion areas are extracted by using two change detection masks and a current frame is directly segmented rather than a difference frame itself so that a good image quality can be obtained at high compression ratios Through the experiments, the sequence of Miss America is reconstructed with visually acceptable quality at the very high compression ratio of 360 : 1

Algorithms for treatment of locally stationary multichannel signals-proposition of a method for change detection based on histograms

Abstract: To take into account slow changes of characteristics in signal which are only locally stationary, an adaptive process for parameter modeling is carried out. An adaptive parametric modeling technique using windows is described. A general process with two thresholds for change detection is discussed. A more general method based on the histograms of the prediction errors is suggested. >

Telephone monitor system

Abstract: PURPOSE:To monitor a change in an object automatically by comparing a variance of a luminance data of a prescribed picture element with a reference value to detect a change in the object. CONSTITUTION:The system consists of a television(TV) camera 1, an object 2, a coding means 3, a luminance data memory 4, a variable calculation means 5, a standard deviation storage means 6, a change detection means 7, a drive control means 8 and a drive means 9. Then a standard deviation data is obtained based on a luminance data of the object 2 in the normal state, and when a change takes place in the object 2, the variance of the luminance data of the corresponding part of the object 2 is changed and the difference is discriminated by comparing the data with the standard deviation data to detect a change in the object 2. Thus, a picture signal obtained from a TV camera is used to monitor a change in the object of the TV camera automatically.

Circuit for detecting changes in vertical and horizontal directions

Abstract: PURPOSE:To obtain a vertical and horizontal direction change detection signal with excellent S/N by applying absolute processing to outputs of plural 2-dimension filters having different pass bands, applying nonlinear conversion to the resulting signal, comparing and synthesizing the result. CONSTITUTION:Plural 2-dimension filters 10-26, 24-28 with different separation bands are provided in each of vertical direction detection circuit and horizontal direction detection circuit to extract each frequency component of an input signal in real time, outputs of each 2-dimension filter are subjected to absolute processing by absolute value processing circuits 18a-18d, 18e-18g and subjected to nonlinear conversion by nonlinear circuits 20a-20d, 20e-20f. A dead band, a tilt and a saturation region in a linear region are set by each nonlinear circuit of each filter and the balance of each detection level is optimized when each filter output is synthesized. Then a synthesis circuit 30 compares and synthesizes each filter output so as to match object control and converts the result to a vertical and horizontal detection signal. Thus, the detection signal with excellent S/N is obtained.

Movement adaptive type luminance signal and chrominance signal separator

Abstract: PURPOSE:To control a pass band width of a variable vertical filter and a variable horizontal filter properly by using an output of a movement detection circuit, an output of a vertical change detection means and an output of a horizontal change detection means so as to designate an address of a storage device thereby reading a prescribed control signal. CONSTITUTION:An input of a control signal generating circuit 24' is an output signal K of a movement detection circuit 11, an output signal of a vertical change detection circuit 20 and an output signal of a horizontal change detection circuit 22. Then an address of a ROM is designated to read control data m'v, m'h stored in advance in response to the state of picture movement and a 2-dimension characteristic of the picture thereby controlling a variable vertical BPF 14 and a variable horizontal BPF 16. Since the vertical and horizontal filter is controlled nonlinearly depending on the output of the movement detection circuit, there is no visual sense of disorder from the moving state of the picture till the standstill state and the filter is controlled smoothly in response to the 2-dimensional characteristic of the picture thereby separating a luminance signal and a chrominance signal accurately.