实现种软件滤波方法及其示例程序（Implementation of a software filtering method and its sample program）.doc

实现种软件滤波方法及其示例程序（Implementation of a software filtering method and its sample program） Implementation of a software filtering method and its sample program 1, limiting filter method (also known as program judgment filter) A and method: Based on experience judgment, determine the maximum deviation allowed for the two sampling (set to A) Each time a new value is detected: If the difference between this value and the previous value is =A, the value is valid If the difference between this value and the previous value is A, then the value is invalid, the value is dropped, and the last value is used instead of the value B, advantages: It can effectively overcome the pulse interference caused by accidental factors C and shortcomings That periodic interference can not be suppressed Poor smoothness 2 median filtering A and method: Continuous sampling N times (N odd number) Arrange the values of the N sub sampling by size Take the middle value as the valid value B, advantages: It can effectively overcome the fluctuation disturbance caused by accidental factors The temperature and level changes slowly, and the measured parameters have good filtering effect C and shortcomings: It is not suitable for fast changing parameters such as flow rate and speed 3 arithmetic average filtering method A and method: N sampling values are continuously taken for arithmetic average operation When the N value is large, the signal smoothness is higher, but the sensitivity is lower N value is small: the signal smoothness is lower, but the sensitivity is higher N value selection: general flow, N=12; pressure: N=4 B, advantages: The utility model is suitable for filtering signals with random disturbance in general In this way, the signal is characterized by an average value that fluctuates near and below a certain range of values C and shortcomings: It is not suitable for real-time measurement with slow measurement or fast data calculation Relatively wasteful, RAM 4, recursive mean filtering (also kn