A code-based approach for labeling in complex irregular regions

Labeling information in a complex irregular region is a useful procedure occurring frequently in sheet metal and the furniture industry which will be beneficial in parts management. A fast code-based labeler (FCBL) is proposed to accomplish this objective in this paper. The region is first discretized, and then encoded by the Freeman encoding technique for providing the 2D regional information by 1D codes with redundancies omitted. We enhance the encoding scheme to make it more suitable for our complex problem. Based on the codes, searching algorithms are designed and can be extended with customized constraints. In addition, by introducing a smart optimal direction estimation, the labeling speed and accuracy of FCBL are significantly improved. Experiments with a large range of real data gained from industrial factories demonstrate the stability and millisecond-level speed of FCBL. The proposed method has been integrated into a shipbuilding CAD system, and plays a very important role in ship parts labeling process.