Aerial Views - Bird's eye perspective, high elevation

[lone chasm]

In art, photography, and cinema, “viewpoint” or “visual perspective” refers to the viewer’s point of view as seen within the artwork. Aerial views, also known as high elevation or bird's eye perspectives, provide dramatic and sweeping visuals in art, photography, and cinema. They achieve directorial goals and challenge typical viewpoints.

The choice of viewpoint can significantly change the impact of a scene. Aerial views create a sense of scale, establish geography, convey themes of isolation or connection, and enhance visual dynamics and movement.

ChatGPT uses your entire conversation to create images, not just short prompts. You can establish a general context and then make specific requests to explore it from different points of view. While DALL-E 3 is not designed to exactly reproduce the same objects, it maintains the general context.

Let’s get three views of a sports event, including two aerial views.

`First prompt: Make an aerial view taken from a blimp flying over a packed football field on a sunny day during the big playoff. The image shows the stadium filled with cheering fans, surrounded by a vast parking lot, capturing the excitement and scale of the event. The news blimp provides cinematic coverage, showcasing the entire stadium and its bustling surroundings from high above.

Second prompt: Now make an image taken by a person sitting in the stadium, looking up at the blimp overhead.

Third prompt: Now make an image from the point of view of a bird flying directly over the field.`

[reef venture]

Prompt created with ChatGPT. Image by Bing image creator: Create a high-resolution, 4-horizontal panel art piece in bright watercolor on canvas, with a bird's eye perspective. Each panel should depict a different view of a city:A city during prehistoric days.A city during early kingdoms.An early city.A modern-day city. Each panel should capture the distinct characteristics and architecture of these eras

[primal island]

While the majority of crop circles are believed to be human-made using simple tools and careful planning, there remains a portion of the population that considers other, more mysterious origins. The creativity and precision involved in making crop circles continue to captivate and intrigue people worldwide.

Aerial view of a minimalistic crop circle design in a wide, expansive field. The crop circle features simple geometric patterns, such as a few large circles and lines, creating a clean and elegant artwork. The surrounding landscape includes lush green fields, distant trees, and a clear blue sky, highlighting the simplicity and beauty of the crop circle. The perspective from a high elevation emphasizes the scale and clarity of the design.

[rare needle]

The Turgot Map of Paris (1734-1739) (and using the principle in historic recreations)

Well before the time of modern air travel, aerial views were used by city planners for various reasons. Aerial views can be very useful in AI when depicting historical events, or landscapes.

The Turgot Map is a highly detailed and accurate Birdseye view of Paris, created between 1734 and 1739. It was commissioned by Michel-Étienne Turgot and executed by cartographer Louis Bretez. The map consists of 21 large sheets and provides an intricate, panoramic view of the city as it appeared in the early 18th century.

Techniques originally used:
Meticulous Ground Surveys: Detailed ground surveys were conducted to collect accurate measurements and observations.
Elevated Observations: Observations from high points like church towers, hills, and later, hot air balloons, provided a broader perspective.
Perspective Drawing: Artists used advanced techniques of perspective drawing to create the illusion of depth and height, making the views appear realistic.

It is public domain, but a very large project. I asked Chat-GPT to let me see a sample visualization, and it provided the following simulation.

Prompt returned: "Here is a detailed segment inspired by the Turgot Map of Paris from the 18th century, showing intricate streets, buildings, and landmarks with a three-dimensional perspective. This historical Birdseye view should highlight the architecture and layout of Paris, with notable features like the Seine River, bridges, and key landmarks such as Notre Dame."
To create a similar type map, you might want to use language like this.* (Image 2, for a 1630 scene of London) “Create a detailed map of London in the 1630s, highlighting notable landmarks, street layouts, and architectural styles typical of that time period, with an emphasis on the historical accuracy and atmosphere of the era.”

[west kernel]

Mapmaker's View There are several aerial perspectives that are used in mapmaking, which result in distinctive visual representations. Planar or orthographic perspective (first image) is from a straight-down, bird’s eye view, often used for large-scale representations. Perspective or pictorial view (second image) is similar to a bird’s eye view but with a more pronounced three-dimensional effect, to highlight landmarks and attractions in a visually engaging way. Satellite view is a top-down perspective from a very high altitude (third image).

Prompt 1: An artistic map of Rome seen from a directly top-down planar perspective, resembling a traditional map. It should include major landmarks like the Colosseum, St. Peter's Basilica, the Roman Forum, and the Tiber River. Use a vintage, hand-drawn style with clear labels and detailed representations of the city's streets and buildings.

Prompt 2: An artistic perspective or pictorial map of Rome, showing the city from an oblique angle with a three-dimensional effect. Highlight major landmarks such as the Colosseum, St. Peter's Basilica, the Roman Forum, and the Tiber River. Use a vibrant, illustrative style with detailed drawings of buildings and natural features.

Prompt 3: “A detailed straight-down satellite map of Rome, showing the city's layout with accurate depictions of buildings, streets, and natural features. Highlight major landmarks such as the Colosseum, St. Peter's Basilica, the Roman Forum, and the Tiber River. Use a realistic style with clear labeling of key areas.”

[west kernel]

Directly top-down view is harder to achieve with ChatGPT/DALL-E than other aerial views. Even when we specifically prompt for a "directly top-down" perspective, resulting images are often in an angled view (as you can see in the first image of my other post). What worked for me was to first create a conversational context by discussing with ChatGPT what the top-down view was, how it was used, etc. After that, I was able to generate truly top-down images (second and third images). In a new chat without a conversational context, the same prompt resulted in the angled view (third image), confirming the efficacy of this approach. Mentioning a "satellite view" and keeping the coverage area smaller (Vatican City as opposed to Rome) also seems to help.

Base prompt: "Create a directly top-down satellite view [photograph/ink and pen drawing] of Vatican City."

[lone chasm]

Aerial views are great for breathtaking nature scenes. In some cases, repeating specific natural features that are complementary with one another may cause the model to produce a larger number of small areas of the image that are attempts to maximize those individual goals. I wanted something showing a rainbow in the mist of a waterfall, viewed from above. This image was produced by providing a very detailed description of a natural scene and letting the system prioritize what it could incorporate. Made with ChatGPT using the 4o model. The output description the system wrote was: "the image visualizing the inspiring scene: amidst rugged cliffs and lush greenery, a magnificent waterfall plunges into a crystal-clear pool below, with the sun casting its golden rays upon the cascading water, forming a vibrant rainbow that bridges the gap between earth and sky. From an elevated vantage point atop the cliffs, this breathtaking spectacle is visible, with the frothy white waterfall contrasting with the surrounding emerald foliage. The pool at the base of the falls mirrors the rainbow, doubling its brilliance, while birds and butterflies flutter near the rainbow’s edge, adding to the magical scene."

Once that image was close to what I wanted, I asked ChatGPT to "use the same image" but change the style in various ways. Since ChatGPT uses the entire conversation, not just short prompts, I asked for a version of an Art Deco travel poster for that location, then progressed to showing a scrapbook memory of that wonderful imaginary trip from many years ago. Here is the revised version: "image showing an open photo album with the 1920s Art Deco travel poster photo, now faded, among other pictures and handwritten notes capturing moments from a trip taken many years ago. The scene evokes a sense of warmth and fond memories."

[unkempt sequoia]

An extreme form of aerial view might be satellite/spaceborne view. Some things can only be truly appreciated with the additional context given by images taken much further from the Earth's surface. For example, many large-scale natural, geological, or anthropological processes and patterns are so vast that traditional imaging techniques are not adequate for capturing their grandeur. Simply adding "satellite view" or "view from space" is generally enough to produce good images, but you might try emphasizing the scale (very small or very large) of certain subjects.

Prompt 1: "Satellite view of a dull olive green-hued algal bloom off the coast of the land. Large beach with river delta. Impasto oil painting style, with thick paint strokes applied by a palette knife."
Prompt 2: "Satellite view of a city at night, the faint lights of cars and buildings clustered in natural blocks. Modern art with faint cubist influences."
Prompt 3: "Satellite view of a rainforest undergoing deforestation. View from the atmosphere. Some of the forest remains, but much of the image is occupied by a barren spread of stumps, with a tiny tractor amid the devastation. Ink sketch illustration."
Prompt 4: "Satellite view of a mountain range. The image is dominated by a wide-angle view of the tall snow-capped peaks. Faded watercolor and colored pencil on textured canvas."