
原文
GPT Image 2 · en
Scene:
A square independent archaeology study composed from a slightly elevated fieldwork viewpoint. Use balanced natural light, crisp evidence hierarchy, and practical site organization so the image feels publishable in a museum report and still visually strong enough for a gallery wall. The frame should explain where excavation is happening without relying on readable markers or diagrams.
Subject:
Show a rediscovery approach scene with survey notes, partially cleared sand at the monument base, and careful inspection of cut channels and facade wear. Include grounded human traces such as ladders, trays, brushes, baskets, light rigs, or measuring tools only when they support archaeological realism and remain visually secondary. The site should stay culturally specific, with architecture, surface treatment, sediment, and environmental context all belonging to the correct civilization rather than a generic dig.
Important details:
Keep the monument recognizably Petra, with faceted rock-cut massing, canyon compression, and carved architectural order that feels cut directly from stone rather than assembled from blocks. Prioritize evidence relationships: exposed layers, cleaned edges, protected surfaces, loose fragments, and transitions between buried and revealed material. Let color, wear, and composition explain what has just been uncovered. The image must feel methodical rather than adventurous. Keep any site tags, notes, or markers abstract and unreadable. Avoid sensational treasure imagery and keep the atmosphere scholarly, patient, and materially convincing.
Use case:
Nabatean Petra Rediscovered single study 5/9.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no modern barriers, no souvenir stalls, no fantasy caravans, no electric lighting, no treasure-chest clichés, no action-movie posing, no arbitrary props, no infographic overlay.使い方
- パラメータを含む元のプロンプト全体をコピーします。
- 被写体、文字、構図を自分の要件に置き換えます。
- 対応する外部ツールで使用し、利用条件と結果を確認します。
開始前の確認
- 参考画像の要否を確認してください。互換モデルの表示は、そのモデルでの生成実績を意味しません。
- 名前、数値、画像内の文字を確認します。モデルの版で結果は変わります。





