It has never been so exciting and easy to combine nature and architecture. I think I’m going to become a hikitect – a cross between a hiker and an architect.

Office work means one thing above all: sitting in front of a computer screen for hours on end. No matter how beautifully you decorate your home office with houseplants, a screen is still a screen.

Artificial intelligence is now increasingly taking over the work of image generation. As a result, we now have more time to spend in nature and be inspired by it.

That would be the ideal, wouldn’t it? Anyone who wants to take up the job of a hikitect, which has just been created with these lines, should go for a walk in the woods during the day with a camera, taking pictures of tree stumps and stones.

When you get home in the evening, the images are transferred to the computer. After a few settings, the AI software takes over and uses the photos to calculate advanced building visualisations for the next construction project.

The self-experiment in nature

Of course, this is all a bit of an over-hyped idea, isn’t it? A self-experiment in the year 2023 shows that this already works in an amazing way.

It’s not a rainy day, the nearest forest is a short drive from home. No other walkers around, just me and my camera. This time I’m not looking into the distance, today it’s the details that interest me. Around tree stumps, around branches, around piles of stones.

As similar as all these structures and shapes seem to be to nature at first, they become more interesting as you spend more time with them, getting closer and closer, looking for that perfect detail with your camera. They could all be human homes, couldn’t they?

From photo to rendering

Why create new forms when many organic buildings already exist in nature? The basis for an hikitect is photos that are transformed into another image using generative artificial intelligence, such as Stable Diffusion.

When converting a real photo into a fictional photo, an essential intermediate step is the creation of a 3D depth model. This model, with its various shades of grey, describes the spatial depth of the image and forms the basis for subsequent calculations.

There are several methods and technologies available to create such a depth model. One of these is already built into Stable Diffusion. If you use the ControlNet module and activate its Depth pre-processor, you can extract the spatial information from the two-dimensional photograph and obtain the depth model required for this method.

From tree stump to residential house

In conjunction with a proper prompt – the textual description of the image to be generated – this depth model can be used as a starting point to generate different variations of buildings. For these images, for example, the following positive prompt was used:

futuristic residential complex with white walls, natural meadow in the foreground, coniferous forest in the background, natural light, photography

Isn’t it fascinating? Of course, the programme doesn’t do static calculations, keep an eye on raw material prices or draw construction plans to the millimetre. But it can help you to overcome mental blocks or make new connections that you might not have discovered on your own.

Of course, all this works not only with wood, but also with stone or other structures and materials found in nature. By changing the prompt, you can easily put the house in a different environment. What about looking at the house at night or at sunrise? Or would a white façade look better than a brightly lit window front?

An architectural visual expedition

With a little imagination and creativity, a trip into nature can become a visual architectural expedition. Of course, nature as a model has been around for a long time. In architecture, for example, bionics is increasingly being used to improve the sustainability of buildings and other structures. By using natural principles and systems, more effective and environmentally friendly designs can be created.

With AI tools like these, however, it is now possible to take first sketches or ideas from the notepad and turn them into photorealistic views at incredible speed. However, this does not replace manual sketching. That is not the aim. It’s more of an aid.

Renderings, completely without rendering software

Brainstorming, thought experiments, idea collection. These are the many benefits of using artificial intelligence as a visualisation tool. The visual results do not require the intermediate step of 3D rendering software and, ideally, save valuable time that would otherwise be spent painstakingly constructing boxes and cylinders.

The new tool opens up new possibilities. Of course, your own creativity and ideas are still needed. Not every image created will show a satisfactory result. But maybe some of them will inspire you to do something new to improve our life in and with nature. I wish you good luck!