Object Dimension Measurement Based on Mask R-CNN
An object dimension measurement system based on Mask R-CNN and monocular vision is introduced to perform non-contact measurement of the two-dimensional size of objects in irregular shape. Firstly, Mask R-CNN is used for detecting all objects to be measured and segmenting each object from the image captured by the camera. Secondly, edge contour extraction is conducted for all object regions and then the minimum bounding rectangle of each object contour can be obtained. Thirdly, according to the result of system calibration, the actual size of each pixel in the image can be acquired. Finally, the actual size of minimum bounding rectangle of objects contour can be calculated. The size of minimum bounding rectangle represents the two-dimensional size of an object. The experimental results show that the object dimension measurement system can accurately and rapidly measure the two-dimensional size of several irregular objects at a time, and the measurement system is robust to the change of ambient light.
KeywordsDimension measurement Mask R-CNN Contour feature
This work is supported by National Natural Science Foundation of China (61603291), Natural Science Basic Research Plan in Shaanxi Province of China (2018JM6057), and Fundamental Research Funds for the Central Universities.
- 1.Wang, Y., Wang, P.F., Yang, Y.W.: Object dimension feature measurement based on image segmentation. Comput. Technol. Dev. 28(2), 191–195 (2018)Google Scholar
- 2.Li, Y.F., Han, X.X., Li, S.Y.: Non-contact dimension measurement of mechanical parts based on image processing. In: International Congress on Image and Signal Processing, pp. 974–978. IEEE, Shenyang (2015)Google Scholar
- 3.Zhang, T., Tang, C., Liu, J.: Bend tube spatial parameter measurement method based on multi-vision. Chin. J. Sci. Instrum. 34(2), 260–267 (2013)Google Scholar
- 5.He, K., Gkioxari, G., Dollár, P., Girshick, R.: Mask R-CNN. In: International Conference on Computer Vision (ICCV), pp. 693–696. IEEE, Venice (2017)Google Scholar
- 7.Long, J., Shelhamer, E., Darrell, T.: Fully convolutional networks for semantic segmentation. IEEE Trans. Pattern Anal. Mach. Intell. 39(4), 640–651 (2014)Google Scholar
- 8.Lin, T.Y., Dollar, P., Girshick, R., He, K., Hariharan, B., Belongie., S.: Feature pyramid networks for object detection. In: Conference on Computer Vision and Pattern Recognition (CVPR), pp. 936–944. IEEE, Honolulu (2017)Google Scholar
- 9.Peng, Q., Song, Y.: Object recognition and localization based on mask R-CNN. J. Tsinghua Univ. (Sci. Technol.) 59(2), 135–141 (2019) Google Scholar