To reduce the ellipticity of the formed pipes

the precise bending technology for forming the steel pipe with JCO process is presented based on plastic bending engineering theory and machine vision technology. By prescribing two-bend in the first bending step

the springback rule can be obtained. With the springback rule and the necessary bend angle

the better punch displacement that determines the forming angle in each bending step is calculated. In the succedent bending steps

the bending process is carried out with the better punch displacement calculated above. And the angle error in each bending step is known by comparing the actual forming angle with the necessary bend angle. By conducting compensation for the last angle error in the next bending step

each precise bending process is realized. In this machine vision system

the image processing and line detecting algorithms suitable for the image of the steel pipe end plane are proposed in this paper. A calibration method is proposed to calculate the exterior parameters of the CCD camera

which employs the equilateral triangle as the calibration board and only captures one image. A mathematical model

which converts the angle in the image into the actual forming angle

is derived. The experimental results showed that the error of measuring the forming angle was less than 0.2°and the ellipticity of the formed pipes was 1.5%. Employing the precise bending technology

the high-quality pipes can be manufactured without the worker’s operating experience.