Sky background-based small target real-time detection and tracking method

Abstract

The invention discloses a sky background-based small target real-time detection and tracking method. The target detection and tracking of a photoelectric tracking system are researched from a multi-scale perspective; and a target detection and tracking strategy for video image real-time processing is put forward according to an embedded system with limited hardware resources. According to the method of the invention, the existence of a target is detected on the basis of a characteristic that the local grayscale features of the target in sub-domains of different scales are different, the existences of a large-size target is detected in real time, and other targets in which main detection objects are small targets are detected through different steps, and therefore, real-time multi-target detection can be realized; according to the imaging features of small targets in a sky background, a small target imaging feature-based weighted centroid tracking method is put forward; and in a tracking process, the tracked targets are scored on the basis of a result obtained after global small target detection is performed, the targets are screened, and finally the real-time detection and trackingof the small targets is achieved. The method of the invention can be used for small targets of visible light and infrared videos.

Description

technical field [0001] The invention belongs to the technical field of information processing, and in particular relates to a method for real-time detection and tracking of small targets based on space and sky backgrounds. Background technique [0002] Since the beginning of the last century, with the development of high-performance models and precision-guided weapons in the military field, precision guidance and air defense have prompted the continuous improvement of target detection and tracking technology. In order to save combat resources and ensure the safety of one's own ground facilities and combat tools, the improvement of target detection and tracking capabilities has become a key technology researched by various countries. The early target detection and tracking technology is mainly based on the radar system design. As all countries in the world are vigorously developing digital and informationized weapons and equipment. The advancement of new combat tools has ma...

AI Technical Summary

Problems solved by technology

The commonly used tracking algorithms are mainly Kalman filter, extended Kalman filter, mean shift, particle filter and other algorithms. These algorithms can improve the tracking accuracy, but nonlinear methods such as particle filter have a large amount of calculation and need to describe the characteristics of the target. Large, complex calculations can not achieve real-time results, and prediction algorithms such as Kalman filtering are not effective for air targets with strong maneuverability and low signal-to-noise ratio

Images