Improved Robust Principal Component Analysis Moving Target Tracking Algorithm and Implementation

In image processing and machine vision, how to accurately and quickly detect and locate the moving target is an important development direction, especially in intelligent robot navigation, unmanned aerial vehicle (UAV) target recognition, and traffic flows control. Machine vision images are high-dimensional and large amount of data, moving target images can be divided into foreground images and background images. For complex scenarios with large amount of image data, this paper proposes an improved robust principal component analysis (RPCA) algorithm and established recursive formulas for solving low-rank matrices and sparse matrices iteratively with augmented Lagrangian functions to improve the speed and accuracy. Further, the kernel paradigm of the low-rank matrix <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\boldsymbol{L}$</tex> is solved using the singular value threshold operator and the paradigm of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\boldsymbol{l}_{1}$</tex> of the sparse matrix <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\boldsymbol{S}$</tex> is solved using the compression operator. The improved RPCA algorithm is tested by real images of conveyor belt moving objects, and the results show that the improved RPCA algorithm proposed in this paper has high accuracy.