Is there a charge for light tilt photography?

I have nothing to do with Twitter, and I found that a great god developed a modeling artifact. Draw a plan in the pixel grid, and then set the relevant parameters, you can generate a 3D hammer model, wonderful …

In fact, Windows 10 system also has a similar fool-like modeling software called "Drawing 3D", which can help you build a 3D model in seconds. For example, if we click on the dog's head in the right column, and then draw the outline and size we need on the panel, a cute 3D dog will appear.

In addition to one-click fool modeling, if you are good at math, you can also build a decent 3D model with basic mathematical formulas.

Recently, VR modeling is quite popular, and even virtual modeling can be directly carried out with only the handle.

As long as you enter the 3D pit, you will find that modeling can lead to Rome. At this stage, the real technical difficulties in the 3D industry are actually model processing and secondary development and application.

This conclusion can also be drawn from the mainstream modeling methods in the industry. There are four mainstream modeling methods in 3D industry, namely polygon modeling, surface modeling, voxel modeling and reverse modeling. These four modeling methods have their own advantages and disadvantages, and also correspond to their own difficulties in model processing.

There is a saying in the industry that "you can model anything with polygons". A polygon is a face formed by a closed path surrounded by multiple sides. As long as you use enough details, you can create any surface. This is also the reason why polygon modeling has become the mainstream modeling method, and there will be no restrictions on making any model.

However, polygon modeling also has disadvantages. It's just an empty shell. Although it can construct the surface of an object, it cannot reflect the internal structure of the real object, and there is still a big gap with the real object.

Moreover, it is easy for the modeler to elaborate too many details, resulting in too many models, so that the baking, rendering and UV display of the models will almost collapse, and the operability of the secondary development will become very low, resulting in an embarrassing situation that the models are completed but not well applied.

Polygon modeling uses straight lines to build planes, while surface modeling is a modeling method using curves to build surfaces, which is very suitable for creating smooth objects. For example, digital products that look silky, cars, and some details of people.

However, the shortcomings of this modeling method are also obvious. One is more troublesome, and the other is difficult to parameterize accurately. Therefore, this modeling method is generally used for visual display to generate videos or pictures.

At the same time, it requires high aesthetics and clarity of maps, and when this model has a large number of high-quality maps, it will lead to a large amount of model data, which is its inevitable shortcoming.

Voxel is actually a three-dimensional pixel block, and the most typical example of voxel modeling is the well-known game My World, in which every object is stacked by hexahedral pixel blocks.

The model constructed in this way, each hexahedron has its own volume, unlike polygon modeling with only one empty shell.

However, voxel modeling has a very obvious disadvantage: too much memory. It is composed of cubes, of course, it needs to occupy the memory of cubes, so this is also the reason why the memory of general voxel games is particularly large, and it may not be portable on general computer equipment.

In the past, modeling methods were based on ideas without physical objects, while reverse modeling was based on physical objects before modeling. Reverse modeling includes a series of related technologies such as photo reverse modeling and 3D scanning reverse modeling. The VR modeling mentioned at the beginning uses reverse modeling.

At present, tilt photography advocated by the state also belongs to reverse modeling, that is, all-round shooting and scanning of real scenes, thus modeling.

The model generated by reverse modeling usually has a large number of faces and graphs, which leads to a huge volume of the whole model, and it is very difficult to even do the most basic display, let alone secondary development.

Based on the inherent shortcomings of the above modeling methods, the model must be reprocessed to meet the different requirements of different scenes for model accuracy, memory, format and so on.

Enterprises need 3D models of cars.

For the automobile manufacturing industry, he will demand very high accuracy of the model, even need accuracy, before it can really be put into production.

If the same car is used in the game, it needs high texture, good fluency and small data to support the picture.

At present, there are generally two ways to deal with the model: one is manual and the other is machine algorithm. Manual operation is complicated, time-consuming and laborious. Probably trying to deal with a large model, it only lost a few megabytes in half a day. In the end, the accuracy was lost, which was not worth the loss.

As long as the machine algorithm technology is excellent, it can achieve twice the result with half the effort and even free hands.

Moreover, after lightweight, not only the accuracy of the model can be guaranteed as much as possible.

Surface reduction and uv development can also be optimized to achieve the best effect suitable for application scenarios.

So if you have a problem with the model processing, you must first find ways to alleviate it, and maybe the problem will be solved.