Autonomous Video Scoring and Dynamic Attitude Measurement

Abstract

The paper focuses on the development and evaluation of an autonomous payload tracking capability for determining time, state and attitude information (TSPI) of all types of airdrop loads. This automated capability of accurately acquiring TSPI data, will reduce the labor time and eliminate man-in-the-loop errors. The paper analyses the problem and then proceeds with the description of the PerceptiVU Target Tracking System (TTS) software adopted for obtaining the TSPI. The key features of this software include a choice of three basic tracking algorithms (dynamic centroid, hottest spot thresholding, dynamic correlation), capability of capturing from both standard analog video sources (such as NTSC and/or RS170) and digital video sources, control of the entire system with an off-the-shelf joystick controller. The paper further describes algorithms to be used in conjunction with the data provided by the TTS to determine system's state variables. A position estimation solution is based on tracking a payload's center (or any other predetermined point) by several cameras with known positions. A pose (position and orientation) estimation solution is based on tracking of four distinctive non-coplanar points. Pre-selected and artificially marked points on the moving target cooperatively serve as beacons, therefore providing precise measurements of the line of sign toward these points. This allows unique position and attitude estimation and no need for additional pattern recognition. In conclusion, the paper provides examples of video data processing and parameters estimation.

Multicriteria Parametrical Identification of the Parafoil-Load Delivery System

“…The data on system's motion is usually obtained either with the help of GPS/IMU unit atop the payload or by tracking the payload using a cinetheodolite system with the following post analysis. 26 As shown by our previous study 4,5 having this data is almost enough for validation of 6-DoF models. For a more fundamental study including separate payload and parafoil behavior (higher fidelity models) and perhaps inflation dynamics, the canopy motion has to be investigated separately.…”

“…Otherwise due to lack of experimental studies and measurements, the parafoil/payload interactions are often postulated in analytical modeling, resulting in theoretical predictions based on uncertain assumptions. Therefore, there is a need to experimentally investigate motion of the parafoil, employing for instance the technique of measuring two angles defining a direction of the riser with respect to the payload, offered by Lee et al 27 , video imaging of the canopy from the camera installed atop payload, 28,29 or by applying the algorithms originally developed to for payload 26 to estimate the pose of the parafoil.…”

Status on Video Data Reduction and Air Delivery Payload Pose Estimation

“…4), we cast this problem as the multivariable optimization problem. Having full information from N≥2 cameras it is needed to find the vector that minimizes the compound functional:…”