The Research Of De-noising Methods In The Automatic And Quantitative Weighing System

With the development of the technology and the massive application of new technology, the automatic online weighing, the fast dynamic weighing and the weighing system have developed greatly. Today, the key point of national development is to develop each kind of automatic weighing system and enhance the accuracy of the dynamic weighing by using new technology further.It often mixed with some interference noises in the signal detection and control system. These noises influence system's stability and precision. These will cause the measurement result to have big error. If the system is expected to have the characteristic of practicality, the noise must be eliminated from the output signal. In the foundation of summarizing de-noising technology in home and abroad, this article has analyzed the noises which exist in automatic and quantitative and proposed de-noising method base on the wavelet transformation.Quantitative weighing system's composition and working principle were introduced. The overall structure was designed, including the selection of sensor, the signal conditioning circuitry, ADC and the selection and comparison of control chip. Then the noises which exist in the overall system were analyzed and summarized in detail.Several commonly used classical de-noising methods were analyzed and compared according to the noises which exist in the system. The performance of these methods was analyzed from different angle, and the existing method's limitation and the existence question were pointed out too. The de-noising method based on the wavelet transformation was proposed on this foundation. And the research has carried on the de-noising using the wavelet transform of the four methods: mold maximum value de-noising method, selective noise filtration, wavelet threshold value method, translation invariant de-noising method. Proposed one soft and hard threshold value compromise method in view of the soft-threshold value function and the hard-threshold value function's limitation. The theoretical analysis and simulation experiment were order to test that wavelet transformation in the de-noising had the superiority.The platform was built for the implementation of the de-noising algorithm based on wavelet in DSP. It includes the design of hardware and software flow. The experiment of de-noising method based on wavelet was carried on aiming at the actual signal in system. The experiment has proven that de-noising based on wavelet can eliminate the noises which exist in automatic weighing system, restore the original signal accurately and improve the precision. These have manifested the superiority of wavelet transformation in the de-noising.