High-Precision Circle Detection for Phase Recognition of Dual-Branch Helical Antenna Element

The dual-branch helical phased array antenna achieves precise microwave beam steering through the mechanical rotation of its antenna elements. Accurate recognition of the phase in a dual-branch helical antenna element (HAE) is critically important for assessing and optimizing the performance of the antenna element. To realize high-precision phase recognition of the antenna elements, this article proposes specific edge detection and circle-fitting algorithms. For the upper end (UE), edge detection with YOLO11n-seg and circle-fitting with the least-squares method are proposed. For the element center (EC), a pixel-based edge detection method and circle-fitting with the random sample consensus (RANSAC) algorithm are proposed. To verify the stability and accuracy of the circle-fitting algorithms, a standard circle dataset for UE and EC (SCUE dataset) is established, and the circle overlap precision (COP) metric is introduced to evaluate the overlap between the standard circles and fit circles. Experimental data analysis shows that the mean COP for UE and EC is 0.951 and 0.807, respectively, indicating a very precise circle-fitting. Statistical analysis of distance between standard circle and fit circle shows that the maximum error of phase recognition is ±3.35°, and there is more than 96% confidence that the error is ±2.11°.