SciVision Vision SDK
OPT's SciVision visual development package: a cross-platform industrial vision SDK covering 2D, 3D, deep-learning inference, measurement, controls, communications, and secondary development APIs.
SciVision视觉开发包 is OPT’s industrial vision SDK for building deployable machine-vision applications around cameras, 2D/3D measurement, deep-learning inference, robot guidance, and production-line communication. It sits below visual workflow software such as Smart3 and exposes the algorithm layer directly for engineering teams that need C++/C#/Python/VB integration.
Product role
SciVision is positioned as a full-function development kit rather than a single algorithm library. The official product page presents it as a platform for 2D vision, 3D vision, and deep learning, with GPU acceleration, multithreading, instruction-set optimization, more than 300 algorithm APIs, and cross-platform support across Windows, MacOS, Ubuntu, CentOS, and Arm.
Architecture
The SDK is organized around an application-development layer, controls and communication modules, and a broad algorithm layer. It is intended to help teams move from proof-of-concept scripts to production applications with reusable UI controls, measurement tools, communication adapters, and sample projects.
Engineering capabilities
The product material highlights fast calls into both interface and algorithm layers, rich example projects, image controls, CPK/GRR controls, TCP/serial/PLC communication, and multi-language bindings. For production use, the practical value is that image acquisition, measurement, inspection, result visualization, and factory communication can be developed within one SDK surface.
Companion book and examples
The companion repository at WillPanSUTD/mvbook contains Machine Vision Algorithms: Principles and Industrial Practice, a bilingual Quarto book with 42 chapters, 9 parts, dual HTML/PDF output, and runnable SciVision SDK C++ examples. The book explains the algorithms independently from the API, then connects them back to SciVision examples where code-level implementation details matter.
