基于SV-DPI的图像坏元修正FPGA自动化验证

为实现红外图像坏元修正FPGA（field programmable gate array）的快速验证，提高测试覆盖性，设计了基于SV-DPI（SystemVerilog-direct programming interface）的FPGA自动化验证平台。采用DPI（direct programming interface）编程接口技术，实现了SystemVerilog平台调用C++编程语言，构建了针对红外图像坏元数据的生成和检测修正模型，建立了两种语言在事务级（transaction level）模型的通信。结果表明相对于传统验证方法，该平台结构简单，可以快速实现激励产生、参考模型构建、测试结果自动比对等功能，实现了红外图像坏元检测与修正FPGA的自动化测试，功能覆盖率达到100%，有效缩短FPGA测试平台搭建和调试周期，提高了测试效率和测试质量。

Automatic Verification of Field Programmable Gate Arrays for Dead Pixel Correction

To accelerate the simulation speed and improve the coverage of verification for a field programmable gate array (FPGA) implemented with dead pixel correction of an infrared image, an FPGA automatic verification platform based on SystemVerilog-Direct programming interface(SV-DPI) was designed. Using DPI programming interface technology, the C++ programming language was invoked by the SV platform. A generator and correction model for dead pixel data of infrared images was built. This established a communication between two languages on the transaction level. The results show that, compared with the traditional verification method, the proposed platform is simple in structure and can quickly generate a test vector, construct a reference model, and check results automatically. It realizes automated verification for an FPGA implemented with dead pixel detection and correction of an infrared image. The function coverage can reach 100%. It effectively shortens the period of construction and debugging for the FPGA verification platform and improves the efficiency and quality of verification.