Research On Data Processing Method Of Automata Movement Based On Image Acquisition

In the test of automatic weapons, the testing of automata motion law is very important content. To facilitate the observation and analysis, we should process the measured data of automata motion law. On the basis of further study on the curve of automata movement and the movement of automaton, we analyze the impact of structural parameters of automatic machines on weapons performance and determine the superiority of firearms performance according to the impact.Through analyzing the characteristics of images which are collected by automata movement parameter testing system based on high-speed non-contact method, spot region prediction method is used to forecast spot areas in the image sequence for determining the scope of the matrix and realize the rapid processing of data. According to the reflection spot image feature points, image processing algorithm is proposed. First, median filtering method is used to eliminate noise; then, edge detection operator is used to realize the image segmentation and regional growth method is used to fill segmentation images; finally, gray center of gravity method of the sub-pixel recognition technology is used to realize the positioning of spot center.At the same time, by analyzing the characteristics of automatic machine movement signal and using low-frequency sinusoidal signal with noise and Matlab software simulates and compares the effect of signal noise reduction of the digital FIR filter, right moving average smoothing, improved wavelet threshold algorithm. For the measured velocity and displacement signals of automatic machine, different methods is used for noise reduction and a threshold improved wavelet-based method of data processing algorithm is proposed. Compared with the test results of Germany OFV5000 series laser lnterferometer speedometer,the noise reduction processing method obtains a better effect and coincides each speed feature points. Many of the measured signal on several different times are tested and the results show that the data processing method has good applicability.