Image Recognition Technology Based On Model-free Adaptive Control Metho

Video image is composed of static background and moving foreground target.Accurate recognition of moving foreground target in video is an important research direction in the field of image recognition.High quality foreground target is the basis of follow-up research such as target tracking and target analysis.The background image in the video can be extracted by using the background subtraction method in the video.Therefore,accurate extraction of background image plays a key role in foreground target detection.There are always various interferences in acquiring background images,such as natural environment changes,camera motion and other factors,which brings serious challenges to accurately extract background images and obtain high-quality foreground targets.The traditional background extraction methods are based on the established mathematical model.After being disturbed by the outside world,the model is prone to distortion,resulting in the low quality of the extracted background image and the incomplete detected foreground target.The method of model free adaptive control(MFAC)only uses the input and output data of the controlled system and does not need to establish the system model,which can avoid some essential problems of the model-based method.This paper systematically studies the background image extraction and moving target detection based on MFAC theory in complex environment.The main contents are as follows:1.Dynamic background video sequence for camera motion.Harris feature detection method is used to determine the coordinate motion relationship of adjacent frame image pixel data,and a MFAC background image extraction algorithm based on compact format dynamic linearization is proposed.The content includes extracting gray background image based on single input single output compact format dynamic linearization MFAC,and extracting color background image based on multiple input multiple output compact format dynamic linearization MFAC.Video sequence simulation results show that in the case of camera motion,compared with the model-based image processing method,the proposed dynamic image processing algorithm based on MFAC theory can stably extract the background image,and the detected foreground target is more accurate.2.Video sequence for rain,snow and bad weather.A background image extraction algorithm based on full format dynamic linearization MFAC is proposed,which includes extracting gray background image based on single input single output full format dynamic linearization MFAC and extracting color background image based on multiple input multiple output full format dynamic linearization MFAC.Video sequence simulation results show that compared with the model-based image processing method,the proposed algorithm based on MFAC theory has better anti-interference ability in bad weather,the extracted background image is more stable,and the detected foreground image is more complete.