1. Introduction Machine vision systems refer to systems that use machine vision products (i.e., image acquisition devices, which are divided into CMOS and CCD types) to convert the captured target into image signals, which are then transmitted to a dedicated image processing system. Based on pixel distribution and information such as brightness and color, these signals are converted into digital signals. The image system performs various calculations on these signals to extract the target's features, and then controls the actions of on-site equipment based on the judgment results. Machine vision has a 15-year history of development. It should be said that as an application system, the functional characteristics of machine vision have gradually improved and developed along with the development of industrial automation. In Europe and America, machine vision has been widely used in industry. Currently in China, with the improvement of supporting infrastructure and the accumulation of technology and capital, various industries are increasingly demanding industrial automation and intelligentization using image and machine vision technologies. Zhuzhou Cemented Carbide Group Co., Ltd.'s profile plant is currently the largest bar production base in China. The Sandvik extrusion press (D-120/250) it imported can extrude various bars: 0.5mm and 1mm bars, double-hole 1mm, double-hole 1.5mm, and double-hole 2mm bars, 7×2.0, 6×2.5, 6×0.4, and 5.5×0.5 tubes, and 3×10 and 1.5×4 flats, etc. The performance of its speed control system is crucial to the quality of the bars; poor speed control can easily lead to breakage or bar stockpiling. Beijing Hollysys Motor has upgraded the extrusion press's speed control system using machine vision and a servo motor-based motion system, significantly improving its performance and thus enhancing the factory's product quality and output. 2. Hardware Introduction The principles of system hardware configuration are as follows: • High sensitivity and wide speed range of the speed control system. • Simple and reliable operation. • Easy system upgrade and expansion. • Self-protection function in emergency situations. • High performance-price ratio. Based on the above overall considerations, the system hardware structure is shown in Figure 1, mainly including: an embedded image processing system with sub-pixel accuracy, an AC servo motor system with excellent fast response, a PLC control system that is flexible in programming and easy to upgrade, an industrial touch screen for easy viewing and modification of system parameters, and an overvoltage and overcurrent protection system. 3. Software Part 3.1 Control Principle The speed regulation system is shown in Figure 2 below. The main purpose of the speed regulation system is to stabilize the extruded bar in a certain area, as shown in Figure 2, that is, to control the edge point of the bar detected by the image processor within the balance zone. In this way, the bar can be extruded normally without being torn or piling up. At the same time, it can also ensure the uniformity of the extruded bar. The extrusion speed of the extruder is greatly affected by the strength of the hydraulic system, the amount of material loaded, the quality of the material, and the thickness of the bar. It is difficult to establish a linear model, and simple PID control is difficult to achieve good control results. In view of the actual situation, we adopted three control methods to adjust the extrusion speed of the extruder: fuzzy modeling of the extruder, differential control, and balance zone control. First, through on-site speed adjustment tests of the extruder, we summarized its fuzzy model—a parabolic model. The closer to the ideal position, the smaller the speed adjustment; the farther from the ideal position, the larger the speed adjustment. Furthermore, the speed adjustment is exponentially related to the difference from the ideal position, as shown in Figure 3. Next, in our on-site tests, we found that in sudden situations, the speed changes drastically, severely affecting the system's stability. We added a differential element to improve system stability. For example, if the extrusion speed suddenly increases due to the hydraulic system, the system requires acceleration. However, according to the parabolic model, the acceleration is far from sufficient to meet the system's needs, leading to material accumulation. Our differential element predicted this situation and accelerated the system accordingly, improving its sensitivity and meeting the requirements. Finally, we adopted a balance zone control. We defined a certain range near the ideal position as the balance zone, assuming the system meets the control requirements within this range. When the bar enters the balance zone from the acceleration or deceleration zone, we adjust the speed in the opposite direction based on the change in edge position. This provides a buffering effect, allowing the bar to remain in the balance zone in the shortest possible time. For example, when the edge point is at the upper boundary of the equilibrium zone (see Figure 2), the speed control system will reduce the speed to make the edge point of the bar descend. When it enters the equilibrium zone, the equilibrium zone control starts to take effect, and the edge position descends. We increase the speed value according to the change in edge position (in our specific case, the change is negative when the edge position descends and positive when the edge position rises. At the same time, the magnitude of the change can reflect the distance of the position change). This prevents the speed from dropping too quickly and overshooting the equilibrium zone, thus causing system oscillation. 3.2 Overall Design The main function of this system is to follow the extrusion speed of the extruder, thereby ensuring the normal operation of the extruder. The software design is divided into three main parts: 1. Machine vision part, which obtains the position information of the extruded bar through an image processor. 2. PLC control part, which writes the control program according to the control principle in Section 3.1. 3. Touch screen part, which allows users to easily view and modify the system parameters and monitor the system operation. 3.3 Vision Part Software Design As shown in Figure 2, the function of finding grayscale edges using image software is used to obtain the position of the extruded bar. The obtained position (digital quantity) is transmitted to the PLC via serial communication at a baud rate of 115200bps. 3.4 PLC Software Design The PLC software design is divided into two main parts: manual mode and automatic mode. In manual mode, the speed control system is manually operated and is not affected by the control laws described in Section 3.1. Speed ​​parameters are modified via the touchscreen to change the operating speed. In automatic mode, the speed control system adjusts the speed according to the control principles described in Section 3.1. Automatic and manual modes can be switched freely. 3.5 Touchscreen Interface Design As shown in Figure 4, the interface design includes four main parts: manual, automatic, parameter viewing, and alarm. Manual: The interface for manual mode. When using manual mode, this interface displays the system's operating status and allows setting the system's operating speed. Automatic: The interface for automatic mode. When using automatic mode, this interface displays the system's operating status and allows modification of the automatic control law's control parameters. Parameter View: Allows viewing of various system operating parameter values. Alarm: When the system malfunctions, an alarm message is given, and some faults can be cleared online. 4. Conclusion This system achieves a perfect combination of motion and vision. It was debugged and put into operation at the profile plant of Zhuzhou Cemented Carbide Group Co., Ltd. The system demonstrated stable operation, simple operation, and speed control that met the performance requirements for actual production.