Robust and accurate moving shadow detection based on multiple features fusion

Abstract In recent years, moving cast shadow detection has been becoming a critical challenge to improve the accuracy of moving object detection in video surveillance. In this paper, we derive a robust moving cast shadow detection method based on multiple features fusion. Firstly, several kinds of features such as intensity, color and texture are extracted sufficiently by means of various measures for the foreground image. Then, the synthetic feature map is generated by linear combination of these features. Consequently, moving cast shadow pixels are distinguished from their moving objects roughly. Finally, spatial adjustment is applied to correct misclassified pixels for acquiring the refined shadow detection result. The effectiveness of our proposed method is evaluated on various scenes. The results demonstrate that the method can achieve high detection rate. In particular, the experiments also indicate that it significantly outperforms several state-of-the-art methods by extensive comparisons.