China
Africa
Explore chapters and articles related to this topic
Computer-Aided Flow Visualization
Published in Wen-Jei Yang, Handbook of Flow Visualization, 2018
A number of useful enhancement techniques employ the process of image differencing (subtraction of one image from another) in order to highlight or remove certain aspects of an image. Recall that subtraction of a filtered image from the original image is employed as a high-pass type of image filter (see Fig. 2b). Other examples of differencing with flow visualization applications are positive–negative addition and sequential scene differencing. Positive–negative addition is a process that has been successfully employed with smoke visualization [15] to enhance the interior details of a complicated concentration gradient. Basically, the positive and the negative of the same image are diagonally displaced a small distance (e.g., 2 pixels) and then added. The resultant image has a “bas relief” appearance, which emphasizes the interior characteristics of the flow structure. An application of this technique is shown in Fig. 5a (note that Figs. 4b and 5a are enhancements of the same original image). Sequential scene differencing involves the subtraction of one image in a temporal sequence from another in order to illustrate the temporal changes in the flow visualization medium. Figure 5b illustrates the effect of subtracting a later sequential image from the dye visualization image shown in Fig. 1a. Note that the effect is to generate highly contrasting regions that reflect the differences in the pictures, which can be quantified.
Characterizing Tropical Forests with Multispectral Imagery
Published in Prasad S. Thenkabail, Land Resources Monitoring, Modeling, and Mapping with Remote Sensing, 2015
E.H. Helmer, Nicholas R. Goodwin, Valéry Gond, Carlos M. Souza, Jr., Gregory P. Asner
e time series method uses temporal change to detect cloud and cloud shadow (Goodwin et al., 2013). It smoothes pixel time series of land surface re¨ectance using minimum and median ¤lters and then locates outliers with multi-temporal image differencing. Seeded region grow is applied to the di¦erence layer using a watershed region grow algorithm to map clusters of change pixels, with clumps smaller than 5 pixels removed to
Dataflow-Based Design and Implementation of Image Processing Applications
Published in Ling Guan, Yifeng He, Sun-Yuan Kung, Multimedia Image and Video Processing, 2012
Chung-Ching Shen, William Plishker, Shuvra S. Bhattacharyya
Image processing on FPGAs is attractive as many interesting applications, for example, in the domains of computer vision, and medical image processing, can now be implemented with high flexibility, relatively low cost, and high performance. This is because many common image processing operations can be mapped naturally onto FPGAs to exploit the inherent parallelism within them. Typical operations for image processing applications include image differencing, registration, and recognition.
A data fusion-based framework to integrate multi-source VGI in an authoritative land use database
Published in International Journal of Digital Earth, 2021
Lanfa Liu, Ana-Maria Olteanu-Raimond, Laurence Jolivet, Arnaud-le Bris, Linda See
Alternatives to post-classification change detection are methods based on the comparison of raw remote sensing data (i.e. optical, DSMs, radar) at two epochs to detect changes where they exhibit important differences. Image differencing is a simple and easy-to-use technique by directly comparing pixel values on imagery obtained from different dates (Muttitanon and Tripathi 2005). However, image radiometry comparison at pixel level often leads to noisy results. Therefore, other raw image comparison methods use texture information instead, e.g. by computing mutual information between images at two epochs over wider windows (Gueguen, Soille, and Pesaresi 2011; Molina et al. 2016). DSM comparison is a specific case among these raw data comparison approaches: it is easier to use in operational situations due to the physical meaning of height differences, and then to define realistic and stable thresholds or to derive change probabilities (Chaabouni-Chouayakh et al. 2010; Guerin, Binet, and Pierrot-Deseilligny 2014; Champion et al. 2010). In addition, the shape of change alerts can be considered.
Image registration for varicose ulcer classification using KNN classifier
Published in International Journal of Computers and Applications, 2018
R. R. Bhavani, G. Wiselin Jiji
In image processing, the process used to determine the differences between two images is Image differencing. In the proposed work the difference between two input images taken at different time intervals for a patient is calculated by finding the difference between each pixel between the source image and target image and generating an image based on the result. For this the two images are first aligned so that corresponding points coincide, and their photometric values are compatible to each other.