In relation extraction tasks， distant supervision is a common method for automatic data labeling. However， this method will introduce a large amount of noisy data， which affects the performance of the model. In order to solve the problem of noisy data， a relation extraction method based on negative training and transfer learning was proposed. Firstly， a noisy data recognition model was trained through negative training method. Then， the noisy data were filtered and relabeled according to the predicted probability value of the sample， Finally， a transfer learning method was used to solve the domain shift problem existing in distant supervision tasks， and the precision and recall of the model were further improved. Based on Thangka culture， a relation extraction dataset with national characteristics was constructed. Experimental results show that the F1 score of the proposed method reaches 91.67%， which is 3.95 percentage points higher than that of SENT （Sentence level distant relation Extraction via Negative Training） method， and is much higher than those of the relation extraction methods based on BERT （Bidirectional Encoder Representations from Transformers）， BiLSTM+ATT（Bi-directional Long Short-Term Memory and Attention）， and PCNN （Piecewise Convolutional Neural Network）.

In this paper, the main idea was to improve the foreground detection method based on statistical model. On one hand, historical maximum probability of which feature vector belongs to background was recorded in the background model, which could improve the matched vectors updating speed and make it blended into the background quickly. On the other hand, a method using spatial feature match was proposed to reduce the shadow effect in the foreground detection. The experimental results show that, compared with the MoG method and Lis statistical model method, the method proposed in this paper has obvious improvement in shadow remove and obscured background restoration of big target object.