MATLAB and HDF

Editor's Notes

• #3: The MathWorks corporate headquarters are located in Natick, Massachusetts, just outside of Boston. In the US, we have field personnel in Detroit to serve our automotive customers, and in California, Washington, and Texas serving customers in aerospace and defense. The MathWorks also has offices throughout Europe, and in Korea. From these locations, The MathWorks offers training and consulting throughout the world Elsewhere, marked with the gray icons, The MathWorks is represented by distributors that represent and support our products in their regions. Note: When appropriate, mention the capabilities of the local representative or one that’s important for the audience (such as Cybernet in Japan for the automotive market) and their close and long-term relationship with The MathWorks. Background: The graphic shows topology (elevation) data, rendered in this map projection using MATLAB and the Mapping Toolbox.
• #4: Our company vision through the use of our tools, is to help engineers and researchers spend less time thinking about the actual programming of their designs allowing them more time to accelerate innovation and creativity in their work We have two flagship products that help with this -MATLAB, flexible programming environment, similar to C -Simulink, graphical user environment for modeling dynamic systems
• #5: Note to presenter: Use this slide to show that we have a number of toolboxes that extend the capabilities of MATLAB and a number of them are useful for Image Processing applications. You do not need to describe all of these toolboxes. Image Acquisition Toolbox Capture images and video from hardware, control devices within MATLAB Database ToolboxExchange data with relational databases Statistics Toolbox Perform statistical analysis, like Principle Components Analysis and K-means clustering Signal Processing Toolbox Analyze one dimensional signals and create filtering kernels that can be used for image processing Mapping Toolbox Analyze geospatial data and place images on map displays using related coordinate data usually found with satellite image files. MATLAB Compiler Deploy components for larger C/C++ projects or deploy stand-alone desktop applications.
• #6: Image data has become a significant part of many applications in scientific fields and engineering activities. Images are captured on a variety of devices at different costs levels from space telescopes and medical imaging systems to webcams and inexpensive digital cameras. As image data becomes more available and useful, it will be involved in more scientific and engineering tasks. The Image Processing Toolbox from the MathWorks provides a comprehensive set of reference-standard algorithms and graphical tools that will help you analyze, process, and visualize image data. Let’s a take a few minutes to explore the different areas of capabilities within the toolbox.
• #7: Note: “Geospatial” is a term that refers to any type of data that is referenced to the Earth. Examples are maps, satellite images, altimetry, topographic maps, and sea surface temperature data. With the Mapping Toolbox and MATLAB, you have an ideal environment in which to perform original research and develop innovative analysis techniques. The Mapping Toolbox provides key functionality to access geospatial data, create 2-D and 3-D map displays, and perform geographic analysis. These capabilities enable you to use geospatial data in MATLAB and take advantage of its well-known capabilities for numerical computation, analysis, visualization, algorithm development and deployment. Recently, we release the first major upgrade to the Mapping Toolbox, which offers improved capabilities in geospatial data access and visualization. The toolbox now supports a broader range of data types, including vector maps, georeferenced imagery and gridded data. In particular, many of you will find it useful that we now support ESRI shapefiles. In the area of visualization, we have completely revamped our map display functionality to incorporate the broader support of data types as well as a brand new interactive mapviewer. Why this is important: Within the past few years, there have been an increasing number of sources for such data. Numerous satellites and airborne systems have come online to generate terrabytes of data. This enormous base of geospatial data is used in a wide range of applications that are not considered “remote sensing” or “mapping.” For example, radar systems engineers are using terrain data to develop better ways of handling ground clutter. Reinsurance companies are using it to model tropical storm risk for asset loss. Earth Scientists are using it to model ocean circulation in the Gulf of Mexico. The toolbox can support a wide range of applications for geospatial data in fields such as defense, intelligence, and homeland security applications, as well as in oceanography, geophysics, and other earth and planetary sciences. Pitch: You may not have thought of using geospatial data to help solve your problem, but if it is related to a specific place on Earth, it could probably help you. The Mapping Toolbox enables you to access this data and use it in the analysis and visualization environment of MATLAB to which you are accustomed.
• #8: As you can see here, we support a wide assortment of geospatial data types. This data falls into three basic categories. The first category is vector map data, which is supported by ESRI shapefiles, Arc ASCII Grid, and a number of well-known data products (I.e. proprietary formats for pre-assembled map data). The second category is georeferenced imagery, which is supported by GEoTIFF and TIFF, JPEG or PNG files with associated world files. The third category is gridded terrain data and bathymetry. MATLAB itself supports several standard geospatial file formats: HDF, HDF5, HDF-EOS, CDF, FITS, and band-interleaved data (note: that last one will be interesting to folks) In addition, the Mapping Toolbox provides built-in atlas and almanac data. This is useful when you need to build a base map to determine your area of interest. It is also useful when political boundaries or coastlines provide improved visualization to your geospatial data. As an example, take a look at the picture on this slide, which overlayed sea surface temperature data (source: MODIS) of the Red Sea on top of a political boundary map of Egypt, Israel, Saudi Arabia, etc.
• #12: For source data, go to www.matlabcentral.com and search on “hdf-eos”
• #14: So, how do we do distributed Computing with ML and SL?? The bottom line is that the distributed computing tools lets you coordinate and execute MATLAB operations on a cluster of computers. A job is a large operation that you need to perform in your ML/SL session. A job is broken down into segments called tasks. You decide how the job is divided up into tasks. A typical job is usually divided into identical tasks, but this is not the only way to define tasks. The MATLAB session in which the job and its tasks are defined is called the client session. Often, this is on the machine where you sit and program MATLAB. IN ML…, in SL… (meaning of a job) The job manager is the software that coordinates the execution of jobs and the evaluation of their tasks. The job manager distributes the tasks for evaluation to remote MATLAB sessions that run in the cluster nodes called workers. The workers execute tasks by calling the function specified by a task, passing the appropriate input data to the function, and then producing a result. The result is then made available for retrieval. Once all tasks for a running job have been assigned to workers, the job manager starts running the next job. Multiple users can send jobs to the same job manager. Each worker is associated with only one job manager
• #15: This slide summarizes the key features of the distributed computing products products Distributed execution of coarse-grained MATLAB and Simulink applications on remote MATLAB sessions See slide #10 Access to single or multiple clusters by single or multiple users See slide #11 Distributed processing on both homogeneous and heterogeneous platforms See slide #12 Support for both synchronous and asynchronous operations Once a user submits a job to a cluster, he/she can wait for the results of the distributed computation (synchronous operation) or can continue working in MATLAB or Simulink (asynchronous operations). While working in asynchronous mode, the user has full control of the toolboxes and blocksets licenses.