Abstract: Systems and methods for automatic detection of disaster relief assets from a satellite image can include a satellite and a processor, which can contain non-transitory written directions that perform a method, which includes the step of defining spectra of interest (SOI). A reference spectral profile that is based on the SOI is built. The reference spectral profile can extend into the near infrared range to eliminate false positive indications. A satellite image of the spatial region of interest (ROI) is taken. For each pixel in the ROI, test spectra of the spatial region can be extracted and compared to the reference spectral profile using a spectral angle mapper (SAM) tool. When the comparison meets predetermined parameters, that pixel of the ROI can be deemed to be indicative of a disaster relief asset, and the geo-reference data for that pixel can be transmitted to a remote user.

Abstract: Systems and methods for determining the speed of a vessel from an overhead image of the vessel's wake can include the initial step of defining a cusp line for the wake. A representative line segment from the defined cusp line that is suitable for building an image intensity profile for the image can be selected. Once the intensity profile is built along the line segment, the wake wavelength ?cusp from the image intensity profile can be found by measuring the distance between successive points of intensity maxima (or successive intensity minima points) along the line segment. Once ?cusp is found, the vessel speed can be determined from the images according to the formula ?ship=?{square root over (1.352*?cusp)}. The systems and methods can be practiced using any overhead imagery systems that display a vessel's wake, including synthetic aperture radar (SAR) and electro-optical (EO) overhead imagery systems.

Abstract: A method involves adapting a vessel detection algorithm based upon received synthetic aperture radar (SAR) image data and associated metadata, detecting one or more vessels within an image tile of the SAR image data by iteratively applying the adapted vessel detection algorithm to successive portions of the image tile, discarding a detected vessel if a false alarm is determined, extracting one or more features from the detected vessels that are not discarded, and generating one or more output products based upon the extracted features.

Abstract: Systems and methods for processing satellite imagery include a satellite, a processor, a database of vessel position data and a computer readable storage medium. The methods process satellite imagery by fusing the imagery with information from the database to automatically identify ships. The methods include the steps of defining an Area of Interest (AOI) and Time of Interest (TOI) for the image and enlarging the AOI according to a time window that brackets the TOI and an assumed vessel maximum speed. Vessel position data from the database for all vessels within the enlarged AOI and the time window is accessed, and fused to imagery position data using Chi-Squared probability analysis. If the analysis meets predetermined probability threshold criteria, the vessel position is assigned to the satellite image to identify the vessel. Otherwise, the operator is alerted that imaged vessels do not correlate to vessel reporting data, or vice versa.

Abstract: A method for automatically detecting and mapping fires based on information extracted from commercial overhead EO/IR imagery, creating geo-referenced files which can be opened in most common geographic information system (GIS) software packages. The method creates a shapefile (*.shp & *.shx) and a Google Earth file (*.kmz) which contain the outlines of the areas from the image being processed with active fire in them, which types of files are typically very small compared to the size of the image file being processed. The method utilizes algorithms designed to process information contained in multi-spectral electro-optical imagery to classify pixels as ‘fire’ or ‘non-fire’. The method also has the ability to identify the approximate length, width, and area of the fires detected.

Abstract: Methods for processing overhead imagery of a vessel include the step of determining an initial classification and classification probability based on the vessel length and length-to-width ratio. Next, mutually exclusive deck features can be extracted from the image. For several embodiments, the extracted deck features that can be spherical tanks, hatches and containers that are stored on deck. The initial classification probability is then weighted using the results of the deck feature extraction step to yield a posterior classification probability for the ship image. If the posterior classification probability is above a predetermined value, the image is assigned a posterior classification. If the posterior probability is below the predetermined value, the vessel image is classified as unknown, and the gross tonnage of the vessel is calculated using the length and width of the vessel.

Abstract: Methods for automatic detection of ships in overhead images of bodies of water are disclosed. The image is initially analyzed to determine if land is present, and the portions of the overhead image where land is present are masked and not processed further. The methods include the steps of chipping the unmasked portions of the overhead image into a series of tiles, discriminating and removing clouds from the tiles using two-dimensional Fourier transforms, and characterizing tile background noise from the water's surface. Different ship detection algorithms are used, according to the level of background noise detected. Detected ships are output into a format that is easily interpreted by the user. The formatted output can also include a confidence rating, or a calculation of the certainty that the detected object in an output file is actually a ship.