Study On Shadow Elimination Method Based On The Properties Of Moving Object Shadow

With the improvement of computer processing speed and camera equipment costs decline, many scientific research achievements in the field of computer vision has been widely applied to the real life of human beings. Widely used and promoted the rapid development of computer vision research and algorithm model was established for many fields or optimization. Humans try to use video equipment and processor to accomplish all kinds of complicated work.This paper studies is computer vision of motion object detection of how to eliminate the influence of the shadows on the moving target. It is well known for the movement of the object, because of the light exposure can produce and target shape similar shadow. The shadows will be as part of the moving target by computer vision detection system come out, because the shadow and target of such a combination, usually shadow and target is together repeatedly, lead to target form of distortion. Another different moving target shadow may be added together, resulting in more complex situation. Computer vision detection system for moving target and shadow separation is a very important question. To solve the problem related to the motion target detection for the next target in accurate location, and target motion and behavior analysis.This paper firstly introduces the simplified model of the shadow, summarizes the characteristics of the shadow. Then, for the shadows of the various attributes were introduced. After for moving target detection and background modeling analysis and research, chose to use a background of the reduction algorithm is gaussian mixture model as the basis of experiment. Then the advantages and disadvantages of all kinds of attribute to spread the detailed research, and through a series of experiments found space transformation and standardization of the method of combining for mild shadow has obvious inhibiting effect, through the improvement to join the suppression coefficient, further improve the effect of reality. In addition, this paper puts forward the fuzzy integral use of various attributes of fusion method, and sums up this year to domestic for dynamic shading general research ideas. For the algorithm of this paper has carried on the experiment shows that good result has been achieved.