Object tracking method based on multi-optical spectrum image sensor

Abstract

The invention discloses a target tracking method with multi-spectrum image sensor fusion. The method adopts a four-level data fusion computing framework with certain versatility. The four-level data fusion computing framework is respectively a multi-characteristic fusion level, a multi-mode fusion level, a multi-model fusion level and a multi-sensor fusion level from top to bottom. The multi-characteristic fusion realizes the integration of a plurality of characteristic clues through the probabilistic method to guarantee enough characteristic information representing targets in different environments; the multi-mode fusion adopts handoff algorithm in a staging mode, integrates two adjacent tracking modes through a certain weighting coefficient, to ensure smooth transition between the modes and guarantee continuous tracking of the targets; the multi-model fusion adopts an interacting multi-model algorithm to solve the problem of stable switching among different moving models of the targets; the multi-sensor fusion adopts distributed type fusion tracking algorithm to select the advantages of all the sensors, and makes up the deficiencies between the sensors, to ensure the tracking to be more robust and more precise.

Description

technical field

[0001] The invention belongs to the field of information technology, can be applied to the fields of intelligent video monitoring, sports video analysis, national defense and the like, and relates to a target tracking method based on a multispectral image sensor. Background technique

[0002] In recent years, with the development of computer vision technology, video-based automatic target tracking technology has been widely used in video surveillance, sports video analysis, national defense and so on. However, in the past, single-sensor tracking systems were constrained by the sensor's own performance limitations, and often could only be applied to the tracking of known targets in certain specific occasions. Due to the influence of uncertain factors such as changing lighting conditions, complex backgrounds, and occlusions, long-term and stable tracking of objects of interest in complex scenes is still a very challenging topic. Therefore, researchers began to...

Examples

Embodiment Construction

[0026] The target tracking method based on the multispectral image sensor of the present invention is carried out in the following manner:

[0027] 1. Build a four-level data fusion framework

[0028] According to the respective characteristics of different sensors, the present invention first constructs a four-level data fusion framework for visual tracking (the structure of the four-level data fusion framework is as follows: figure 1 shown), including: from bottom to top are multi-feature fusion level, multi-mode fusion level, multi-model fusion level, and multi-sensor fusion level. The multi-feature fusion level uses a probabilistic method to achieve fusion based on multiple visual cues; the multi-mode fusion level uses a staged mode switching algorithm to combine two adjacent tracking modes with a certain weighting coefficient to make a smooth transition between modes; the multi-model fusion level It is realized by using interactive multi-model (IMM) algorithm; the multi-...

technical field

[0001] The invention belongs to the field of information technology, can be applied to the fields of intelligent video monitoring, sports video analysis, national defense and the like, and relates to a target tracking method based on a multispectral image sensor. Background technique

[0002] In recent years, with the development of computer vision technology, video-based automatic target tracking technology has been widely used in video surveillance, sports video analysis, national defense and so on. However, in the past, single-sensor tracking systems were constrained by the sensor's own performance limitations, and often could only be applied to the tracking of known targets in certain specific occasions. Due to the influence of uncertain factors such as changing lighting conditions, complex backgrounds, and occlusions, long-term and stable tracking of objects of interest in complex scenes is still a very challenging topic. Therefore, researchers began to...

Examples

Embodiment Construction

[0026] The target tracking method based on the multispectral image sensor of the present invention is carried out in the following manner:

[0027] 1. Build a four-level data fusion framework

[0028] According to the respective characteristics of different sensors, the present invention first constructs a four-level data fusion framework for visual tracking (the structure of the four-level data fusion framework is as follows: figure 1 shown), including: from bottom to top are multi-feature fusion level, multi-mode fusion level, multi-model fusion level, and multi-sensor fusion level. The multi-feature fusion level uses a probabilistic method to achieve fusion based on multiple visual cues; the multi-mode fusion level uses a staged mode switching algorithm to combine two adjacent tracking modes with a certain weighting coefficient to make a smooth transition between modes; the multi-model fusion level It is realized by using interactive multi-model (IMM) algorithm; the multi-...

Images