为了探究小功率激光模具自动修复技术, 利用同轴视觉采集系统采集模具的裂纹图像, 结合数字图像形态学细化处理识别裂纹位置信息, 建立了数字图像处理流程, 得到裂纹的轨迹信息, 将裂纹轨迹信息矢量化后,经曲线拟合生成数控代码,导入到数控系统完成激光模具修复。裂纹图像经图像去噪增强、形态学细化等处理后, 能够有效地得到裂纹中心线, 将裂纹位图矢量化后转为DXF文件格式,通过CAM软件生成数控加工代码。结果表明, 该方法加工精度达到0.0368mm,满足模具修复的精度要求; 通过图像形态学细化处理技术可以实现激光模具自动修复。这对激光加工设备实现自动化和智能化提供了理论支持和技术基础。

In order to repair dies automatically with low-power laser, crack images were taken by a coaxial vision acquisition system. Crack position information was acquired combining with digital image morphological thinning processing technology. Digital images processing were established and crack trajectory information was obtained. And then, dies were repaired by computer numerical control (CNC) system from numerical control (NC) code generated from curve fitting vector image. Crack center line was effectively obtained after image denoising, enhancement and morphological thinning treatments. NC codes were generated with the help of CAM software after crack bitmaps were converted to DXF file format. The results show that repair precision of dies can reach 0.0368mm and meet the repair demands of dies. Dies can be repaired automatically by means of image morphological thinning processing. It is theoretical support and technical foundation for automation and intelligence of laser processing equipment.