基于机器视觉的风口回旋区温度检测算法

 应用机器视觉理论，引用微元积分思想建立了辐射图像信息与高炉风口回旋区二维温度场梯度分布的关系模型，并针对CCD非接触测量回旋区温度过程中因温度过高易出现CCD输出过饱和电流问题，提出了大于CCD噪声总量和防止饱和电流输出的快门时间控制模型。通过小型煤气燃烧实验和高炉风口现场实验表明：该温度检测算法正确，快门控制模型能有效地防止图像出现“白炽”现象，提高了温度测量峰值。

Raceway Section Temperature Measurement Algorithm Based on Machine Vision

The relational model based on machine vision between two-dimension temperature field grads distribution of the blast furnace raceway section and radiation image information is established by the calculus ideas. Due to the problem of CCD supersaturation current output in the process of CCD nowcontact raceway high temperature measurement, so the CCD shutter control model is also provided to prevent from the supersaturation current output and exceed the CCD noise signal. The results of experiments on a mini-type coal gas furnace and practice test show that the algorithm is accurate and credible, at the same time, the shutter control model, which is effective for avoiding image incandescence phenomena, improves temperature measurement range.