The invention can realize online automatic counting. It adopts the technical path of video acquisition, image processing, multi-target real-time identification, multi-target real-time tracking, multi-level alignment fault-tolerant algorithm and multi-direction operation control, takes the international advanced machine vision technology as the principle and the original patented algorithm as the core, and counts the bars moving with the chain bed in real time.
The end face image of moving bar is collected from one side of the steel moving machine, and real-time counting is realized after image detection, image recognition and target tracking. When the bundle count value reaches the set value, the steel moving machine is controlled to start and stop. It can be used to display the tapping position in real time and guide workers to tap manually.
System accuracy:
The counting error of bars above 1 and φ 14 is less than 0. 1 ‰? .
2. The counting error of bars below φ14 is less than 0.2‰.
3. Allow multi-layer stack counting.
4. Practical technology development of batch shearing process.
5. Support the measurement and packaging of rated and set quantities.
6. Communication interface with other systems.