Abstract:

Experimental analyses were conducted on the color mixing effect of the developed full⁃color FDM 3D printers， verifying the color mixing uniformity of the designed mixing nozzles. Thus， an RGB⁃CMYKW color matching calibration algorithm was proposed based on FDM 3D printing of polylactic acid（PLA） filament and BP neural network. CMYW four⁃color silk materials were mixed in different proportions to obtain 529 mixed color samples. Using SOFV⁃1xi image acquisition equipment to extract color （RGB） information from mixed color samples under standard D50 light source， and combining the color information extracted from the standard color card under the same acquisition conditions for color correction， 529 RGB-CMYW color conversion sub samples were obtained. The sub samples were used for training to determine the FDM 3D printing full⁃color matching calibration algorithm. The developed full⁃color FDM 3D printer was used to print 24 color chips， and the corresponding color differences among the printed color chips and the standard 24 color card were analyzed and calculated. Results indicate that the color error difference is generally small and the color features may be well replicated. At the same time， actual 3D color model design and printing are carried out， further verifying the reliability and practicality of the proposed color matching algorithm.

Key words: fused deposition modeling（FDM） 3D printing, color matching, color mixing, calibration algorithm, neural network

摘要：

对所开发的全彩色熔融沉积建模（FDM）3D打印机混色效果进行试验分析，验证了所设计的混色喷头的混色均匀性。在此基础上提出了基于FDM 3D打印聚乳酸（PLA）丝材和BP神经网络的RGB⁃CMYKW颜色匹配标定算法。将CMYW四色丝材按照不同比例混合，获得529种混色样片。利用SOFV⁃1xi图像采集设备在标准D50光源下对混色样片进行图像颜色（RGB）信息提取，同时结合标准色卡对同一采集条件下提取的信息进行颜色校正，得到529个RGB-CMYW的颜色转换子样。对子样进行训练，确定了FDM 3D打印全彩色颜色匹配标定算法。利用所开发的全彩色FDM 3D打印机进行24色片打印，分析计算打印色片与标准24色卡的对应色差，结果表明，色差普遍较小，能较好地复制颜色特征。同时进行了实际的三维彩色模型设计及打印，进一步验证了所提出的颜色匹配算法的可靠性和实用性。

关键词: 熔融沉积建模3D打印, 颜色匹配, 混色, 标定算法, 神经网络