MVTec designs MVTec HALCON for high performance on modern hardware. HALCON automatically exploits multi-core and multi-processor systems and supports GPU as well as NPU acceleration.
MVTec provides Automatic Operator Parallelization (AOP), which parallelizes suitable operations automatically without requiring code changes. This enables performance gains on multicore systems with minimal effort.
For advanced scenarios, HALCON also supports explicit parallel programming and thread-safe execution, including within the development environment HDevelop.
MVTec provides HDevelop as an integrated development environment tailored to machine vision workflows. HDevelop supports interactive development, visualization, debugging, and export to production languages such as C++, C#, and .NET.
With HDevEngine, MVTec enables direct execution of HDevelop programs inside applications without compilation. This supports flexible architectures and rapid iteration in productive systems.
MVTec designs MVTec HALCON to make optimal use of available hardware. HALCON supports modern instruction sets such as NEON, SSE2, AVX2, and AVX512, as well as GPU acceleration.
With the AI Accelerator Interface (AI²), HALCON supports dedicated AI inference hardware. By abstracting models from specific devices, AI² enables future-proof deployment on both PC-based and embedded systems.
MVTec provides MVTec HALCON with advanced machine vision technologies to support robust inspection, reliable recognition, and accurate measurement across demanding industrial scenarios. MVTec combines deep learning, anomaly detection, matching, 3D vision, morphology, and robust reading of bar codes and 2D codes into one consistent software environment, so teams can select the right approach for each task without changing toolchains.
MVTec focuses with MVTec HALCON on efficient and transparent workflows for developers. Interactive development with HDevelop (IDE), clear visualization, and consistent data handling reduce complexity in daily work.
Deep-learning-based applications can be implemented efficiently using integrated workflows and tools such as the MVTec Deep Learning Tool. Custom operators, camera calibration, and application-specific extensions can be integrated seamlessly.
