Edge detection is a fundamental task in image processing. Traditional methods like Sobel and Canny operators perform well in simple scenarios but struggle with high noise, high resolution, and complex backgrounds, often leading to edge fragmentation, false edges, and parameter sensitivity. In response to these challenges, quantum computing-based edge detection algorithms have garnered significant attention. One notable approach, Q-Sobel, adapts the classical Sobel operator into the quantum computing domain by utilizing quantum circuits for edge extraction. However, despite its innovative approach, Q-Sobel still suffers from issues such as parameter sensitivity, circuit noise, and limited adaptability to varying image textures. To address these limitations, this paper proposes an improved quantum edge detection algorithm, Parameterized Optimized Quantum Edge Detection (POQED), which introduces a rotation angle scaling factor (scale) and an adaptive measurement strategy (shots) to dynamically optimize the quantum circuit. By integrating multi-scale Sobel preprocessing, POQED leverages both quantum and classical advantages for enhanced edge detection. Experimental results, averaged over multiple parameter settings and test images, demonstrate that POQED achieves superior edge continuity (F1-score = 0.92), outperforming the classical Canny algorithm (F1-score = 0.76) by 21%, and significantly improving over Q-Sobel (F1-score = 0.79). Furthermore, POQED exhibits better noise robustness, achieving a PSNR of 10.5 dB (compared to 7.3 dB for Q-Sobel) and a lower MSE of 0.12 (versus 0.15 for Q-Sobel). These improvements highlight the potential of POQED in practical quantum image processing applications, offering a more robust and adaptive approach to quantum-enhanced edge detection.