GeoSpace is a combination of pen interaction and voice interaction of three-dimensional geometry rendering system. In the Unity3D engine through a separate design of the hidden, rotating line surface and other three-dimensional graphics algorithm, the production of this multi-layer module based on the MVC architecture to achieve rotation, measurement, display and other functions of the three-dimensional geometry software. In addition, we also design and train handwriting recognition and speech recognition modules, which are integrated into the system through interface calls, and realize the multi-modal interaction of geometry teaching software for the first time.

The system can be divided into three modules according to functional categories: geometry module, command module and identification module.

Geometry module is the basic module of the system, which realizes the data processing and logical behavior of solid geometry. It is mainly responsible for adding, deleting and storing geometric points, edges, faces and other elements, rendering three-dimensional graphic objects, controlling the position and perspective transformation of the lens, and drawing depth maps for calculating virtual and real lines of edges.

The command module is responsible for responding to the user's operation on the geometry and UI view, modifying the geometry model according to the operation content, and notifying the geometry view and interface to update in response.

Recognition module is the core module of the system. Based on the traditional graphical interface interaction, the system integrates the pen interaction of blackboard writing and the voice interaction of simulated lectures, and adopts multi-modal interaction to enhance the operability and practicability of the system in the actual teaching environment.