Improves pipeline inspections
Our terrain ground vehicle autonomously identifies buried pipelines and its depth. Combined with Aerial monitoring and inspection using UAV it autonomously navigates pipelines, It identifies pipeline defects in real-time and geo-tags the musing a high accuracy GPS. Non-contact sensors and AI
technology are introduced to overcome shortcomings of conventional intrusive inspection tools. Artificial Intelligence (AI) and Machine Learning (ML) are used for the inspection and management of the buried metallic pipes. Cracks, corrosion and erosion can be detected and preemptively addressed with proper maintenance.
Enabling object tracking through scenes with large light intensity changes
Visual surveillance of dynamic scenes is critically important for a broad range of applications. Architecture for real-time extraction of of maximally stable extended extremal regions (X-MSERs) has been developed. The system is comprised of a communication interface and processing circuitry configured in hardware to receive data streams of an intensity image and a depth image in real-time. The system determines X-MSER ellipses parameters based upon the strong extremal regions and X-MSER criteria.
Improves seismic data acquisition and processing
Underwater seismic acquisition is a standard industry process to collect marine data. The developed method enables the processing of multi-component marine seismic data in order to estimate the properties of the seafloor and sensor calibration filters in a shallow water environment. The method applies a time-frequency-wave number transform on the acquired seismic sea bottom data and applies a time varying frequency-wave number filtering process to accurately process the seismic data.
Enhancing data security in cloud computing
Critical security and data privacy issues can exist if the dynamic random-access memory (DRAM) content is not sanitized or wiped out before being allocated to a newly provisioned virtual machine in cloud computing. Various methods have been developed to erase volatile memory that is comprised of erasure circuitry that prevents refreshment of the memory upon occurrence of a predefined event by the randomization of data stored, by manipulating reference voltage, and by controlling refreshment of stored data.
Improving object detection and tracking from video streams
Visual surveillance of dynamic scenes is critically important for applications that range from computer vision to robotics to national security. Architecture for linear-time extraction of maximally stable extremal regions (MSERs) has been developed. The system is comprised of image memory, heap memory, a pointer array and processing hardware. The system can determine MSER ellipses based on the range of components and MSER criteria.
Improving the vision of computers and robots
Extracting affine-invariant regions and features from image data is used in numerous computer vision and robotic applications. The developed output detection and tracking method involves the receipt of depth image data corresponding to a depth image view point relative to objects detected. With the method, a series of binary threshold depth images are formed from the depth image data to detect and/or track the object.
Achieving real-time video analysis
Object detection, recognition, and tracking from a video stream is critically important for applications that range from computer vision to robotics to national security. Hardware architecture for real-time extraction of maximally stable extremal regions (MSERs) has been developed to improve image extraction. The system is comprised of a communication interface and processing circuitry configured in hardware to receive data streams and extract MSERs in real-time.
Effective and reliable hands-free control of a computer mouse
Use of a computer is challenging for some individuals with physical disabilities due to the inability to manipulate the computer mouse. The developed computer eye controlled computer mouse interface could enable manipulation of the mouse without requiring sophisticated hardware. The interface is comprised of a camera, a display, memory, and a processor with data acquisition, calibration function, eye detection, face detection, and gaze mapping functionality.
Enhancing real-time video analysis
Object detection, recognition, and tracking from a video stream is critically important for applications that range from computer vision to robotics to national security. Architecture for real-time parallel detection and extraction of maximally stable extremal regions (MSERs) has been developed. The system is comprised of a communication interface and processing circuitry configured in hardware to receive data streams and extract MSERs in real-time.
Reducing the impact of seizures in epilepsy patients
The occurrence of random and sudden seizures in epilepsy patients can be stressful and potentially dangerous. The machine-based patient-specific seizure classification system could detect seizure precursors and potentially suppress the seizures. The system-on-chip (SoC) includes a hardware-efficient log-linear engine to realize non-linear support vector machine-based seizure detection. Also, the system enables multi-channel electroencephalography data acquisition and storage.
Eliminates manual and inaccurate data extraction
Charts and graphs are frequently utilized by scientists and engineers to present and interpret data. The systems and methods developed enable data extraction from charts and graphs to provide a usable format for further data analysis/interpretation.