A Sequential Detection Method of Abnormal Behavior Based on Multi-factor Inconsistency Metrics

Abstract

The invention discloses a method for sequential detection of abnormal behavior based on multi-factor inconsistent measurement. This method fully considers the position, speed and movement direction information of the target, and realizes real-time abnormal detection of abnormal behavior of the target through online learning and sequential abnormal detection. It specifically includes the following steps: 1. Define input and output variables; 2. Initialization; 3. Repeat the corresponding anomaly detection for each data point in the test sample and each sample in the training sample sequence; 4. After the abnormal detection is completed for each data point of the current test sample, update the training sample sequence; 5. Update the multi-factor Hausdorff distance matrix; 6. The updated training sample sequence and the updated multi-factor Hausdorff distance matrix are used as new input variables to perform anomaly detection on the next test sample. The parameter setting of this method is simple, the false alarm rate is controllable, the accuracy of anomaly detection is high, the project is easy to realize, and it has broad application prospects in the field of early warning and monitoring.

Description

technical field [0001] The invention relates to anomaly detection technology in data mining and high-level fusion technology in information fusion, and belongs to the field of pattern recognition and intelligent information processing. Background technique [0002] With the continuous improvement of information fusion theory and the wide application of information fusion technology, the intelligence processing system can automatically or semi-automatically complete the detection, tracking, track correlation and attribute judgment of targets through the fusion process of detection level, position level and attribute level , forming a continuous and stable target track. With the continuous increase of target types and numbers and the continuous improvement of the performance of early warning and monitoring systems, more and more target intelligence data are formed and exist in various early warning and monitoring systems. How to let the computer automatically discover the abn...

AI Technical Summary

Problems solved by technology

However, these methods generally have problems such as complex parameter settings, inaccurate statistical models, ineffective control of false alarm rates, and poor online learning effects.

Images