An infrared target detection and tracking method

Abstract

The invention provides an infrared target detection and tracking method which comprises the following steps: selecting density, rectangularity and geometric invariant moment as static variables, and establishing a static decision criterion; And selecting a target contour area, a perimeter, a speed, a self-adaptive segmentation threshold and an ROI position as dynamic variables, and establishing adynamic decision criterion. a target static variable and a part of dynamic characteristic variable values are calculated by adopting a first-frame target detection algorithm; A subsequent frame firstly utilizes an improved local adaptive threshold segmentation algorithm to segment an image, then utilizes static and dynamic decision criteria to screen out segmented real targets, and finally calculates and updates a to-be-tested target characteristic parameter set. The method has good robustness for complex background update, scene change, and scale, gray scale and contour feature change of thetarget.

Description

technical field [0001] The invention belongs to the technical field of target detection and tracking, in particular to an infrared target detection and tracking method. Background technique [0002] Target detection and tracking based on infrared imaging has been widely used in precision guidance, battlefield monitoring, unmanned reconnaissance, visual navigation and intelligent monitoring. In such applications, the complexity of dynamic scenes and target image backgrounds is the main factor affecting the robustness of target detection and tracking algorithms. How to overcome their interference is an important factor to be considered in algorithm design. For infrared precision guidance, when the target is imaged on the infrared detector from a few pixels to fill the entire field of view, the scene update speed is fast, and the coupling between the target and the background is strong. Especially at the end of the guidance, the scale and grayscale of the target itself And the...

AI Technical Summary

Problems solved by technology

Among them, the first type of method can be better used for detection and tracking tasks, and it is relatively easy to implement and has strong real-time performance. However, the algorithm design is difficult, and it is impossible to establish a deeper and stable feature appearance model of the target.

The second type of method improves the accuracy and robustness of tracking, but the establishment of the appearance model needs to rely on a large number of training and test data sets, and the algorithm operation requires high hardware support

At present, the embedded processors of infrared precision guided weapons mainly include ARM, DSP, ARM+DSP and FPGA+DSP. These processors have insufficient performance in floating-point operations, matrix multiplication, convolution operations, and memory bandwidth. The transplantation of complex model algorithms and real-time processing requirements of the second method

Images