步骤 2 —— 舞台（GPT Image 2），然后绘制你的镜头路径

Step 2 — The stage (GPT Image 2), then draw your camera path

Generate a top-down championship-night stadium. Then literally DRAW the camera route on it — red line, 4 numbered nodes: stands → dive → orbit the duel → goal.

That hand-drawn line becomes your camera track.

Prompt 👇

A photorealistic aerial drone photograph of a packed football stadium at night, captured from high above at a steep top-down angle. The full pitch fills the center of the frame and stays perfectly readable — crisp white line markings on immaculate green grass with clean mowing stripes — while the surrounding grandstands remain visible around all four edges of the frame in steep perspective.

# ATMOSPHERE — CHAMPIONSHIP FINAL NIGHT
World Cup final night energy: stadium floodlights at full power making the pitch glow against the darker bowl of stands, tens of thousands of fans forming a dense sea of color, thousands of phone flashlights sparkling like stars throughout the crowd. One end of the stadium is a solid block of red-clad supporters, the opposite end a solid block of deep-navy-blue-clad supporters, the two long sides mixed. A light stadium haze hangs in the air catching the floodlight beams, and subtle golden confetti drifts over the stands only — none of it on the pitch, the grass stays completely clean.

# PITCH
Empty pitch, no players, no staff, no equipment — pristine grass ready for kickoff. Sharp complete white markings: center circle, halfway line, both penalty areas and goal areas, corner arcs, and both goals with white nets clearly visible.

# CAMERA & RENDERING
High-altitude drone shot, wide-angle with minimal lens distortion, tack-sharp focus across the entire pitch, broadcast-photography color treatment, realistic HDR night exposure — bright saturated glowing pitch, moody dark-blue night sky just visible at the frame corners. No text, no logos, no scoreboard graphics, no watermarks.