In the industrial production, high temperature steam pipelines are usually used to transport high temperature steam and high temperature industrial waste water, etc. But the pipelines are often placed in complex environments for safety, which is not conducive to the inspection and maintenance of pipelines for the workers. How to quickly locate the location of the steam pipeline in the complex background and distinguish the surrounding environment has become an urgent problem to be solved. Because the Otsu algorithm cannot meet the requirements mentioned above, an improved Otsu algorithm is proposed based on the cellular immunity mechanism. According to the characteristics of the pipelines and the complex background in the infrared images, the algorithm calculates two different thresholds which can be used to extract the pipelines and distinguish the complex background. With the help of the Quartus Ⅱ software， a hardware and software system based on FPGA has been built to realize the data communication and transmission, and the improved Otsu algorithm has been verified. The experimental results show that the algorithm can be applied to the infrared images of the pipelines to obtain better effect. Compared with several edge detection operators and the classical Otsu algorithm, the improved algorithm has a higher true positive rate (TPR) and a lower false positive rate (FPR), both for the pipelines extraction and the complex background distinction.