Research On Data Fusion System Of Plaster Ointment Thickness Measurement

In plaster factory, manual operation method is usually adopted to control quality of a plaster ointment. First, the plaster is marked, it passes through a machine and becomes dry, a sample of standard size is obtained by a quality inspector. Second, the quality inspector weights the sample with an electronic scale. Then, the weight of the plaster ointment is controlled through than workers adjust the machine to control the thickness of the plaster ointment according to measurement result. This method has a lot of shortcomings: (1) real-time property isn't well. (2) Measurement error is big. There are many error sources during the manual operation process, such as different sampled point, size of sample and worker's reading data etc. These factors make random errors exist in the measurement result data. (3) Human cost is high. Some special people are needed for sampling and weighting. This method can't satisfy the requirement of modern production. Since the thickness and the weight of plaster ointment have some function relationship, developing a real-time system to measure the thickness or the weight of plaster ointment automatically is needed to improve production quality and factory effective. In this dissertation, data fusion's theory and method are the main content. Two thickness measurement systems based on laser method and capacitance micrometer method are combined together. A general scheme of data fusion for plaster ointment's thickness measurement system is proposed. In this dissertation, a new laser measurement method is put forward. Linear laser beam is used as light source. Linear photocell is used as detector. This method can expand measurement area and improve measurement accuracy. In this dissertation, the thickness measurement system of capacitance micrometer is also researched. Mechanical part of the capacitance micrometer system is designed. In this dissertation, the data fusion's theory is studied and several kinds of data fusion's methods are introduced. Based on the data fusion's theory, a kind of BP neural network model is found. The data fusion for the measurement results of the two systems is realized. A BP algorithm is put forward. The algorithm is written by C++ compute language in VC++6.0 design environment. The data fusion'result compares with the measurement result of each kind of the two systems. The compared result expresses that the measurement accuracy is enhanced.