Remnant detection method fusing multiple models

Abstract

The invention discloses a remnant detection method fusing multiple models, and relates to the technical field of remnant detection. In the remnant detection method fusing multiple models, a suspectedremnant target is tracked by using a tracking technology; relevant parameters can be adjusted according to actual conditions; when related targets are lost due to pedestrian flow shielding in the scene, continuous tracking of the targets can be effectively ensured; a background modeling result and a target detection result are effectively combined, and false detection targets are greatly reduced through multiple filtering condition screening. According to the method, the foreground and background separation model and the target detection model are logically and ingeniously fused, the advantages of the foreground and background separation model and the advantages of the target detection model are combined, and therefore the defects are better overcome, high robustness can still be kept forcomplex light and shadow environments and various kinds of shielding under the condition that the system performance is guaranteed, and high accuracy is guaranteed.

Description

Technical field [0001] The invention relates to the technical field of remnant detection, in particular to a remnant detection method integrating multiple models. Background technique [0002] With the rapid development of the national economy, traditional manual-based video surveillance can no longer meet people's needs for security. Intelligent video surveillance has received more and more attention and applications as a more efficient and convenient means. In subway stations, high-speed rail stations, airports, shopping malls and other crowded public areas, the emergence of unidentified leftovers that are deliberately wrapped (such as flammable and explosive materials, toxic chemicals and other dangerous goods) is currently the most worthy of close attention. Concerned about a major hidden danger that threatens public safety. Therefore, real-time detection of unidentified remains has become an important application in intelligent video surveillance. [0003] Most of the existi...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the deficiencies of the prior art, the present invention provides a remnant detection method that integrates multiple models, which solves the problem of low accuracy in the current traditional remnant detection method, which cannot cope with complex scenes, and is susceptible to various occlusions, light, Problems affected by factors such as shadow changes

Images