1. Track Detection Algorithm Based on Trace Correlation Using Hough Transform
- Author
-
A. A. Monakov
- Subjects
trajectory processing ,track detection ,hough transform ,Electronics ,TK7800-8360 - Abstract
Introduction. Track detection is one of the main tasks to be solved in trajectory processing. This task can be efficiently solved using the Hough Transform. A track is considered detected if the number of position measurements received in a number of consecutive radar scans and falling into the same cell of the parameter space (accumulator) has exceeded the detection threshold. However, the effective practical application of the Hough transform requires a sufficiently long time of measurement. Under a small number of scans given for track detection, measurements are also accumulated in those accumulator cells where their traces intersect. Therefore, in order to detect true tracks, additional processing is required to distinguish measurement clusters from different targets based on their geometric proximity. In addition, a large amount of memory and computational operations for the accumulator maintenance significantly increase the computation load of the trajectory processor.Aim. To design a simple and false-detection resilient algorithm for detecting tracks without the Hough accumulator in the processor memory.Materials and methods. In the proposed algorithm, the construction of measurement traces in the Hough accumulator followed by selection of cells with the largest number of traces passed through them is replaced by computation of the cross correlations of the traces and clustering of measurements based on the maximum similarity of their traces.Results. Mathematical simulation with the scenario parameters selected in the paper confirmed the accuracy of the proposed algorithm in detecting all tracks existing in the radar field of view and its efficiency in conducting error free association of target position measurements.Conclusion. A false-detection resilient algorithm for track detection was created based on the Hough transform. The algorithm does not require the Hough accumulator in the processor memory.
- Published
- 2023
- Full Text
- View/download PDF