The Vision
Cone Penetrometer
(VisCPT) 
umseal.jpg (44836 bytes)

Read about the VisCPT below or visit our newest sites (April, 2000): 

VisCPT video clips of soil liquefaction

 

VisCPT log from Treasure Island

 

Still images from Treasure Island


rclip1.jpg (18333 bytes)

The Vision Cone Penetrometer (VisCPT) was developed at the University of Michigan in 1995/96.

rclip2.jpg (15692 bytes)

By adapting miniature video cameras to the electronic cone penetrometer, the VisCPT has overcome the major previous shortcoming of the Cone Penetration Test (CPT). Namely, the inability to visually observe the soil.

rclip3.jpg (14768 bytes)

Click on Dr. Raschke to see the VisCPT viscpt1t.jpg (5494 bytes)

rclip4.jpg (16956 bytes)

TheVisCPT consists of two cameras, lenses and lighting systems with individual housing units. The two cameras record the soil images through synthetic sapphire windows.

rclip5.jpg (15994 bytes)

Click to see the camera assemblies   viscpt4t.jpg (7167 bytes)

rclip6.jpg (21386 bytes)

    viscpt2t.jpg (7330 bytes)  Click to see the illuminated sapphire window

rclip7.jpg (21845 bytes) Each camera system operates at a different level of magnification thus providing fields of view between 2 mm and 20 mm (diagonal).
rclip8.jpg (24417 bytes)

The images are recorded continuously in real time as the probe is advanced at the standard CPT rate of 2 cm/sec.

rclip9.jpg (20550 bytes)
The sequence of  VisCPT images shown at left reveals soil grain sizes ranging from medium sand near the top to silty fine sand at the bottom.

  
rclip10.jpg (26082 bytes)

   Horizontal field of view       =     1.0  cm  
   Total vertical length            =   14.2  cm

rclip11.jpg (21182 bytes)

A field-portable S-VHS recording system and monitors allow for immediate field inspection as well as recording for later viewing.

rclip12.jpg (22268 bytes)

Click for the recording system viscpt3t.jpg (4256 bytes)

rclip13.jpg (21810 bytes) After recording, the continuous stream of images is digitized.  Computer vision image-texture techniques analyze the data for soil grain size.
rclip14.jpg (23139 bytes)

coming soon:

Links to Ali Ghalib's image analysis techniques and Seung Cheol Shin's neural networks

rclip15.jpg (22510 bytes)

 


In summary, the VisCPT offers the following advantages to geotechnical and geoenvironmental site characterization:

rclip16.jpg (22393 bytes)

Visual inspection of soil without sampling, a particularly attractive feature for geoenvironmental applications.

A permanent visual record of the soil is collected, eliminating the need for sample storage.

rclip17.jpg (23077 bytes)

Ambiguous CPT data may be resolved. Exact elevations of stratigraphic transitions are revealed.

Continuous recording of the soil stratigraphic column at very high resolution. Anomalous clay lenses, fissures and sand seams, even as thin as 0.5 cm. are detected.

rclip18.jpg (22367 bytes)

Images are digitized and computer vision techniques are used for image-texture analysis of the soil.

Grain size distributions can be inferred from the textural analyses.

rclip20.jpg (22293 bytes)

Suggested Reading:

Hryciw, R.D., Ghalib, A. M. and Raschke, S. A. (1998) "In-Situ Soil  Characterization by VisCPT", Proceedings of the First International Conference on Site Characterization (ISC'98), Balkema Press, pp.1081-1086.

rclip21.jpg (20884 bytes)

Raschke, S.A. and Hryciw, R. D. (1997) "Vision Cone Penetrometer (VisCPT) for Direct Subsurface Soil Observation", ASCE Journal of Geotechnical and Geoenvironmental Engineering, Vol. 123, No. 11, pp. 1074-1076.

Raschke, S.A., Hryciw, R.D. and Ghalib, A.M. (1997) "Adding Vision to the Cone Penetration Test", Proc. of the 6-th Great Lakes Conference on geotechnical and Geoenvironmental Engineering, Ann Arbor, MI pp. 37-50.

rclip22.jpg (23021 bytes)

 

Related Reading:

Ghalib, A.M., Hryciw, R.D. and Shin, S.-C. (1988) "Image Textural Analysis and Neural Network for the Characterization of Uniform Soil", accepted for publication and presentation at the 1998 ASCE Computing Congress.

rclip23.jpg (24067 bytes)

Hryciw, R.D., Ghalib, A.M. and Raschke, S.A. (1998) "Methods for Soil Characterization from Images of Grain Assemblies", Proceedings of the 2nd International Conference on Imaging Technologies: Techniques and Civil Engineering Applications", ASCE, pp. 88-99.

Hryciw, R.D., Raschke, S.A. and Ghalib, A.M. (1997) "Characterization of Particulate Assemblies Through Computer Vision", in Selected Research in Environmental Quality, Proc. of a Joint USAF/Army Contractor/Grantee Meeting, Jan. 14-17, Panama City, FL. , pp. 131-136.

rclip24.jpg (24774 bytes)

Hryciw, R.D. and Raschke, S.A. and (1996) "Development of Computer Vision technique for In-Situ Soil Characterization" in Emerging Technologies in Geotechnical Engineering, Transportation Research Record No. 1526, pp. 86-97.

Raschke, S.A. and Hryciw, R. D. (1997) "Soil Grain Size Distribution by Computer Vision" ASTM Geotechnical Testing Journal, Vol. 20, No. 4, pp. 433-422.


rclip25.jpg (24037 bytes)

For additional information contact the VisCPT Group:

  Prof. Roman D. Hryciw      Professor Scott A. Raschke
  University of Michigan       Lehigh University
  romanh@umich.edu                     sara@lehigh.edu

© 1998 Roman D. Hryciw