Bar automatic counting system as an important part in the bar rolling mill production line. And it is the key technology to realize negative deviation rolling of bar production, delivery according to theoretical weight, and international market. Also, bar automatic counting system is an important technology to improve the automation level of bar production.

Bar automatic counting system is a complex Mechatronics system, its functions include automatic counting and automatic separation. The automatic counting mainly calculates the number of rods passing through and gives the position information of the rods when the accumulative value approaches the set value. The automatic separation is to arrange the bar into a single layer parallel to each other when the accumulative value of the bar approaches the set value, and to separate automatically when the accumulative value of the bar reaches the set value.

Firstly, the photoelectric sensor (including image sensor) is used to sense the information of the bar’s moving position and get the number of rods passing by. Considering the minimum overlap of the projection area at the end of the bar, the image sensor should be placed at the end of the bar according to the placing principle of the image sensor. Then the number of bars in a single frame image is identified by static image processing method, and the number of bars in the image sequence is calculated by dynamic image processing method.

Secondly, we should consider installing paving and finishing devices in the process of bar transmission, and arrange the bar laying state into a single layer parallel to each other.

After finishing, there may be a small amount of cross-cutting phenomenon in the bars. Alarm should be given after the cross-cutting bars are found, and manual finishing should be carried out.

Finally, the ultimate goal of this system is to use visual servo technology to control the manipulator action when counting to the set value, and to separate the bar automatically. The image is regarded as feedback of closed-loop control, and the manipulator is controlled by image moment technology to reach the target position in the image. In the end, the qualified bar is separated from the next bundle of bars.

Considering that the aim of online automatic bar counting is automatic bundling, it is necessary to control the manipulator to automatically separate the front bundles from the back bundles when the number of bars reaches the set value.

Therefore, when the counting approached the set value, the mechanical device was started to lay the bar flat; when the counting reached the set value, the transmission chain was stopped and the manipulator was controlled to move in that position to realize automatic bundling.