Research On Axis Extraction Methods For Target’s Attitude Measurement

Attitude parameters are target’s main parameters in shooting range. The information of pitching and yaw angle plays an important role in real-time control and post-processing. There are lots of attitude measurement techniques. Optical measurement is one of the most important ways. Generally, if the target is an axis symmetry object, we can extract the axis in digital images which are got by slow-motion cameras, and calculate the three-dimensional information of the target’s axis by binocular intersection to achieve the attitude information, such as pitching angle and yaw angle. However, the factors of yawp, dimension and illumination can influence the result seriously in practical situation. If the algorithm is not appropriate, measurement’s accuracy cannot be ensured. The article’s main work is to conclude and improve different kinds of axis extraction methods of symmetry target, add new algorithms for some given situation, and give the reference to use certain method for specific situation. The work has high application value for completing attitude measurement mission efficiently and accurately.Firstly the article introduces the composing of system’s hardware and software, as well as the binocular intersection theory. By analyzing the two main parts of error generation, we get the conclusion that the precision of axis extraction decides the attitude measurement error, and the relationship between them.Classify the different situations in axis extraction mission, recapitulate the two main approaches, which are based on the detection of edge-line and the calculation of area’s inertia, and give the whole process for each direction. Analyze algorithms in each step of axis extraction. Compare the advantages and disadvantages of GAC contour extraction and edge detection algorithms, which is completed by gradient modulus operators and Canny operator. Compare the applicability of line detection by Hough transform and LSD, give the conception of minimum length and angle acceptance to improve the Hough transform. Analyze the theory of area’s inertia and its applicable condition, and add a new algorithm for illumination condition.Based on the summarization and improvement, 5 different algorithm has been supplied for various axis extraction situations, and experiments has been done to ensure the accuracy.The article has researched the axis extraction methods in different conditions, and will greatly improved the the system’s accuracy and adaptation capability.