Study On Measurement Of Restoring Force In NFAL Based On Image Recognition

Near-Field Acoustic Levitation(NFAL) is a kind of levitation phenomenon which based on high energy density ultrasonic waves, it can levitate piece above the ultrasonic transformer. NFAL has plenty of advantages, for example:high capacity of supporting,high stability of levitation,less requirement for environment. NFAL is used widely in some situations because of the non-contact feature. For NFAL, when floating plate shifts, a restoring force comes to being and pulls the plate back to balance position. The restoring force is usually very small, which is about several milli-Newton (mN). For general measure methods, it is difficult to get the certain value of restoring force. This paper conducts a research, which focuses on the measurement of restoring force, on the basis of Newton’s Second law as follow:In order to automatically recognize and measure the target, a special marker, which is composed of two concentric circles, is attached to the floating plate. Before the analysis and measurement, a video of the marker’s motion process is taken, and then analyze every frames of the video by MATLAB. Because of the particularities of the marker, six criterions were extracted for the following recognition and analysis. Besides, a kind of simplified region growing method which is faster and more efficient was proposed. This method is useful in some particular situations though the recognition accuracy is relatively lower.This paper adopts several sub-pixel location methods for the location of the maker, and then selects the best one as the final method. At the same time, a lot of simulated recognize experiments were taken for some main factors, after which the statistic data were obtained. These statistic data can be used to analyze and compensate the system error.The position data of the marker was analyzed and fitted by MATLAB, then velocity curve and acceleration curve could be got easily by differential operation. In order to verify the accuracy of measured data, the reason of making error and the error propagation are analyzed carefully in all processes. At last, a conclusion is drawn that the measurement accuracy of restoring force is about ±0.25mN at the confidence of 95%.This paper designed and completed all related programs. The program can automatically analysis and calculate the restoring force after importing a certain motion video of the maker. A lot of unnecessary operation in restoring force measuring was cut down because of the automatically-analyze program, it is easy and efficient for using.