Study And Application In Target Recognition Of Mixed Configuration Data Fusion Algorithms Based On Rough Set Theory

With the rapid development of technology, various information systems which face to complex appliance background appear, in which battlefield direct system,robot system and target detect track recognition system etc. are representations. The kind and quantity of sensors in these systems are great increased, and the requirement to sensor performance is also enhanced. Each kind of sensor has its own character and limitation, it can not offer accurate information at any time. Multi-sensor data fusion technology can synthetize every side information from multi sensors, can acquire more complete and more accurate information of observation object, then acquire more accurate and more rapid decision and estimation.Firstly, this paper expounds the basic theory,basic concept and the character,merit,apply background of multi-sensor data fusion, summarizes the level and model configuration form of data fusion, summarizes and classifies the fusion arithmetic in common use. Secondly, this paper presents the rough sets theory concept, and proposes using rough sets theory in data fusion to recognize, and this method can improve the fusion effect. To multi-sensor data fusion algorithm based on rough sets theory and its appliance in target recognition, via clearing up and summing up research literature in existence, this paper does deeper research and does simulation experiment. Meanwhile, this paper expounds the data fusion methods based on D-S evidence and neural network, researches the appliance of these algorithm in target recognition.With people's requirement to multi-sensor data fusion system performance becomes higher and higher, it can not achieve the requirement only rely on one fusion method. So, this paper attempts use mixed configuration data fusion algorithm to realize the best fusion effect, also analyses and compares the advantage and disadvantage among a few target recognition fusion algorithms. At last, through comparing each fusion method, this paper presents a kind mixed configuration fusion algorithm based on rough sets theory,neural network and D-S evidence theory, and uses this method in target recognition. The simulation experiment result proves that this method offer an efficient approach to improve fusion recognition effect.