Section Half Imaging Logger ( SHIL )

X401

User's mishandling during cropping generated image file issue

• At the beginning of the expedition, some images were not uploaded to LIMS and .roi files ended in the Error folder. Scientists had to image a couple of times to get the image uploaded in LIMS (and so in LIVE). This issue seemed to happen randomly:

  • The image (.jpg) was not created and only two image files (.tif and _trim.jpg) are saved in the IN>IMAGE folder (see below).

Open

•  

  •  The .roi file went into the Error folder as expected. See below the usual error message given by MUT2.
    
    

    Open

    

  • When opening the image files created by IMS, it appears that:

    • The .tif image saved is actually just the black section label (name, text_ID)

    • The _trim.jpg is in fact the image normally saved as .jpg image. The cropped image is not generated/saved and therefore the image saved as _trim.jpg is the one with ruler and label.

  • There was no image displayed in LIVE and the image printed on the VCDs was the image with virtual label and ruler (see picture below).

Communication from Tim Blaisdell, developer, regarding the issue (December 26, 2023):

It's pretty clear that what's wrong is that IMS is not writing one of the JPGs -- the uncropped one in this case. That's really strange because the cropped and uncropped JPGs are written right in the same bit of LabVIEW code -- the uncropped one first.   And the cropped image is created just before that from the uncropped one -- so you'd think if there was something wrong with the image, so that it fails to write the uncropped image, it would also fail to write the cropped one.  It's weird that it would fail to write the uncropped image, then succeed in writing the cropped one.

I have added code to the SHIL so that immediately after it finishes writing the cropped and uncropped JPGs, it verifies that the files exist.  If either file does not exist, a message box pops up saying which file was not created, and what the expected path was.
The message ends with "Notify an IODP tech or developer of this immediately!"

• The issue with the SHIL continues after Tim's changes, and the window set by Tim bugged and never popped up, unfortunately. Scientists mentioned later on that they sometimes clicked twice on the CROP button. We tested it and we could reproduce the SHIL "issue" every time we did click twice.
  
  

  • What SHOULD happen: when the crop area is selected, users when they are happy click "CROP" and then click "SAVE" to save the image and/or print the VCD. If the user is not happy, the user can "UNCROP" or "DISCARD" the image by clicking on the corresponding buttons.

  • What IS HAPPENING: The user clicks "CROP" once and the button does not become disabled. The cropped image appears on IMS as it should be. For some reason, the user clicks a second time on "CROP" before saving the image and the cropped image displayed changes to the image with ruler and virtual label. The "CROP" button becomes disabled at that time and the user SAVES the image after the second click. This leads to two image files created (one tif containing the virtual label, and one _trim.jpg which shows the .jpg image with ruler and virtual label) instead of three. The printed VCD will have the image with ruler and virtual label.

• SOLUTION: Tim Blaisdell fixed the issue and the CROP and UNCROP buttons are immediately disabled when clicked on. It should be impossible to double-click either of these buttons.

• A new VCD template was created to account for a customized description printed in A3 size.

X400

• The SHIL was making a weird sound when going back to home position/load after imaging a section. This happened towards the end of the expedition and only at higher speed. Screws were tighten (a couple were lose) and the worm drive lubricated. Sound stopped and SHIL is working fine. 

X398P

• IMS software updated

X398

• Scientist presenting problems setting the correct length for IW and other samples taken on the bottom of the section. A recommendation could be to add the length of the section to the information that appears in the scream when the label is scanned.

• For further information please visit Physical Properties Lab Notebook or Imaging Specialist Report.

X397

• The Newall encoder was dirty and caused vibrations while the track was moving, cleaned the shaft and the Newall encoder.

X397T

No Issues

X397P

No Issues

X393

• Dry and wet images were taken for hard rock cores. Dry images were selected to be displayed as the main image in LIVE.

X390

• Temperature controller on light system fail at the beginning of the expedition, ET specialist Garrik changed the temperature control and add two releases to both heating system and an external release is connected to buzz alarm to cut the general power to the heating system, wiring is explained in ET report.

X392

• The hardware belonging to the old light system was removed from the track.

• SHIL Track was adapted physically and IMS was reviewed in order to make more space to work with the 360 image adapter. It was moved 4 cm forward.
  
  

• The alignment 'stop' on bottom of SHLF and 360 core trays are half-lapped to keep them from slipping. Now both trays lock in the same spot in the bottom rail. To remove the tray first move it forward (towards the camera) and then lift.
  
  

• An existing strip of PC fans were placed along the track to help dry section halves. This will be removed at end of expedition. This is kept above the DHML printer.

• Few times during the expedition the camera lights didn't come ON. The didn't come ON neither using the Utility. Every time the problem was solved by restarting IMS.

• Hard-rock curated cores were 360 imaged, however no 360 composites were done. EPM confirmed the composites would be completed onshore if they choose to do so.

Software 

• Duplicate ROI and RGB files were being created when a crop was discarded (see programmer report). The issue is resolved. 

• Issue of WRND sample info not displaying (noticed by previous expedition) was fixed by Bill Mills. 

X391

• January 20th,2022- SHIL right hand side temperature controller failed.  The alarm triggered at 50 degrees C.  The heating element was constantly applying heat to the system.  A new temperature controller was installed by the SIEM electricians.  Technical staff input the same settings into the temperature controller as the previous controller.  Settings for the controller were documented and placed in the lab notebook.  The old controller was placed in the drawer and marked "Failed 391, further testing required."  The temperature controller is now on the PP check out sheet as item number PP5024.

• Change made to SHIL IMS program to facilitate the upload of the high resolution RGB data.

  • In the ImageRGBDataWriter.vi, the RGB High Res Global path and file component name "rgb_high_res" were added to the Section and Control Data Paths. This results in the file path for the high resolution data file being written to the "FILE" section of the RGB file.

  • The GetSingleData.vi was edited to remove the RGB High Resolution file path from the LSIMG file.

  • The High Resolution RGB data is available in the Expanded RGB report. The file extension is rgb.dat.

X396T

• Light system was changed: New frame was installed, circular LEDs were maintained and a lens was added at the bottom of the lights to help defuse the light on the core. A warming/cooling controller to maintain an operational temperature of 30 degrees C was installed. Several calibrations and tests were performed after the changes to assure the correct behavior of the instrument.

  Nov 19, 2021 One of the two controllers (left hand side or FWD controller) used for maintaining the temperature failed and the heater heated up the lights to 150 deg C. We do not think any damage was caused to the lights. New controllers and new heaters are added to the requisition (heaters with less wattage). ET Jurie has the controller working again, but if it failed once it may fail again. See ET tech report 396T for details.

• A Temperature alarm was install on the track, it will advise the user when the light temperature reaches 50 degrees. 

• If, when performing the calibration, the color standard squares appear on the image as rectangles, it indicates that the camera is not at the correct height. Go to INSTRUMENTS > Camera: JAI Camera Setup and by moving the camera up and down, reach on the Profile box the pixel value of 200 pixels per centimeter.

• See Imaging Specialist report and ET's report for further information.

X396

• Experienced issues with the camera losing focus during imaging.  This resulted in batches of cores needing to be reimaged.  The cause was undetermined.  It may have been from someone bumping the camera or from the track movement shaking the camera.

• Scientists struggled to get the core at a correct wet/dry stage for imaging.  Scientists were often lightly dabbing the rocks as the SHIL was imaging to get the images to look wet.  The issues appeared to be related to increased reflection with the bright lights on the wet surface.  

• Dark sediment cores had to be left to dry longer before imaging to avoid getting reflected light 'streaks' in the image.

• A very common issue was the instrument software losing connection with the LIMS database. The 'Connect to database' button would often 'fix' the issue but only for one image and then it would be disconnected again. The manual tab could be used and lengths entered manually for each image taken but this was cumbersome for the scientists to do for every section.  This issue has been on going for many expeditions.

X395C

• Review Imager Specialist report.

• In order to allowed more space to the technician to manipulate the 360 imaging gadget the load position of the SHIL was moved.

• NOTE: For hard rock imaging be sure that the core is completely dry (360 and section half images).

• It was noticed by users that each time that Windows updates were installed the motion of the track become slower.

• The positioning plate of the lower core tray was moved to aligning the section with the 360 widget, in order to have a better control when the pieces move from one to another.

• Sarah Kachovich 3D printed the aft lateral-protector of the camera, used to imaging whole round sections. She also added rubber covers to both sides of the camera. The protectors used for section halves are storage on PP-2B drawer.

• Two new standards that could be used for calibrate or test the SHIL camera were done by Sarah Kachovich.

X395E

• Experienced issues with the calibration of the camera and vertical streaks in the images.

• To see camera calibration issues and suggestions, please refer to the Imaging Technical Report here.

• The QP card had some sediment on it and was scratched over the period of operation. The Imaging Specialist replaced it. We added a note to SHIL work stations stating the fact that the QP card needs to be kept pristine in order to post-process images if needed, or it needs to be swapped to a new card.

• The SHIL computer was running very slow at times. The switch from the “scan sample” screen to the “crop screen” took several seconds, in the meantime the screen stayed white. This would cause concerns if scientists would run the track.

• Unresolved bugs still exist and are listed on the software bugs page in Confluence.

X395P

No Issues

X390R

• Two RGB extractions methods were written in R software (Lisa C. method) and MATLAB (Alejandro A. method) both permits the extraction the RGB values and plots, this is two alternatives to brute force RGB extraction using magic, notepad, and excel. See more in  RGB extraction methods in the Lab Notebook.

• Many tests were carried out with light angle variation and light distance to core, more details in Image and Microscope There are 2 configurations now available onboard, the old 70-degree setup and a new 45-degree setup.  The brackets, baffle, and light shields for the 45-degree setup are kept together in PP2B drawer.

• No decision has been made on which configuration is “best” therefore the SHIL was restored to the configuration it has been used in since installation with the 70-degree light angle.

See Imaging Specialist X390R Expedition Notes and SHIL report X390R for further information.

X390C

For more information about the SHIL review Imaging Technician report.

• Nov 23, 2020 The CANCEL button does not function properly in (at least) the Image Correction tab of IMS on the SHIL computer. It acts as a save action button. Problem was solved for the Image Correction page.

• IMPORTANT: There are different versions of QP cards onboard. Technician will be aware of which version is using because they could have different Dark Gery, Mid Gray or White values. A label with the version of the QP card on use was set on the SHIL.

• Oct 10, 2020 The track began to get stuck when moving and displayed the E-Stop window. The M-Drive motor seized, and we ended up replacing it for new one. The worm screw of the track was greased.

• Oct 10, 2020 Issues found during X384 and 390P (Red and green bands, vertical stripes) were found to be largely related to calibration, since then the occurrence of artifacts has been reduced. A file with the calibration images was saved on IODP_Share\PhysProps\390C SHIL Calibration-Fixed GreenRed streaks.

• Oct 25, 2020 QSG was reviewed and updated in Confluence.

• Oct 26, 2020 The light stand was moved across, to better center the illuminate over the core.

Testing

• Sarah Kachovich and Alexis Armstrong performed several tests during this expedition. Tests were conducted in two ways:

  • Reviewing RGB values on several QP cards measurements. A drift on the blue spectra was detected in high intensity values. It related with increment of the lights not stabilizing during measurements. The data shown that after approx. 10 mins, the LEDS stabilized, but the heat sink is not efficient enough to run the SHIL at this level. Moreover, the calibration needs to be done for the correct stabilization (or temperature as proxy) time. This drift in RGBs intensity values could cause problems when imaging really white cores or highly reflective/wet sections. See Imaging Specialist report for more information.

  • Measuring real cores. Three types of tests were realized. (1) Duplicate measured sections were analyzed, (2) measuring the same section continuously during multiple runs, and (3) measuring a complete core, seven sections, continuously during multiple runs. In all cases, the RGB data produced by the SHIL were compared. A drift on red, green and blue intensities could be seen as the lights stabilized (as a proxy, the temperature of the lights is increasing). A drift in the blue intensity diverges at a faster rate than in red or green intensity over time. After 7–10 minutes, the red, green and blue intensities were observed to reach an equilibrium and should be used as guide when running the SHIL, until we understand this phenomenon more.  Overall drift values were small, with the highest value of blue drift corresponding to ~5. Change in RGB values between duplicate sections (sections which were re-run for higher image quality) were on average ~3. We do not know if this drift value of 5 is significant; however, to decrease blue drift, running the instrument before taking your image is shown to have a significant impact, as shown in the figure below. Data related with this experiments could be find on: T:\IODP_Share\PhysProps\Testing and Calibration Notes\SHIL\390C SHIL testing.

X390P II

• During 384 the technicians noted issues with the core images included a green and red band on the photos and vertical striping down the core sections. During 390P further testing was conducted to determine the cause of these artifacts in the images.
  No solution was found during 390P.

• Issues noted during testing included:
  A drop off in the images (starboard side) is noted during the calibration process. The lights are not centered over the core, causing this darker area in the photos. In order to center them, adjustments need to be made to the position of the track or the position of the camera.

• Vertical stripes noted in the 384 images have been attributed to our calibration methods. During calibration we de-focus the lens and we apply pixel corrections. If these corrections are not done during the calibration process, the vertical stripes are not seen. The current thinking is that the grey card we are using is not completely uniform and these slight differences are causing an issue.

• A new standard grey card that is less susceptible to the high heat of the lights arrived in the air freight and will hopefully be more uniform and not have the issues with the vertical striping.

• The green and red bands seem to become less after the cameras were swapped, but further testing is needed.

• A loose connection was found in the power cable for the SHIL JAI camera. Photo below illustrates the connection that was repaired by the ET.

• The cameras were swapped on Sept. 29th. Camera #91148 was taken off the SHIL and Camera #91171 was installed. This swap was done to determine if the artifacts seen in the images (green/red bands) were caused by the camera itself.

• Recommend testing both cameras with the new grey card and full calibration during 390C. Camera #91148 may have had an incorrect setting that was causing issues with the images.

X384/390P I

• Physical Camera settings when I first came on: F-stop set at f/22 and distance between bottom of LED mount and core surface 5.1cm. I changed the f-stop to f/16 but did not move the lights down (in Bill Mills instructions it was stated that lights should be between 2-4cm away from core surface) as that would leave ~ 1/2 cm clearance between light shield and core surface.

• Calibration performed following instructions from Bill Mills. Took this operator multiple tries to get the calibration image looking good. During calibration had an issue keeping the cardboard gray cards flat as they would flex and warp under the heat. It was recommended not use the plastic cards as they could start melt under the lights.

• Tech report notes: Wet images of hard rock seemed to be more difficult with more intense reflections off of the core surface.

• During calibration was supposed to adjust the pixel delta value. I was unable to adjust this value. Unclear at this point if this is user error or a bug in the software. Comparing past pictures of the setup window to current settings, the pixel delta value has remained the same since at least X378.

• At the beginning of the expedition had issues with IMS disconnecting with the network every time a section was run. The issue was only reported on the SHIL and stopped occurring after several days.

• The gray scale card at the top of core images shows a faint green line (below). After conversations with the imaging specialist, it is believed this is an artifact from the camera and has been seen in previous expeditions. A new camera is in process of being ordered.

• More green lines were noticed. One looks to be about the width of a pixel and extends down the length of the core. There are two thicker more diffuse green lines in the center and edge of the image.

• A new light shield was 3D printed to better accommodate 360° imaging. The current one doesn't allow enough room to take metal strip off.

• Test Images were taken for the Geology LWG to evaluate the new lighting system. Five section halves were each measured under eight testing conditions, at two speeds for each test. Section halves were two dark play cores, one light play core, the concrete core, and a section of Styrofoam and rags. Images were uploaded to LORE under 999 U9999A 11H1 through 11H5. Images and pictures of the lighting system were copied into data1 to be transferred to shore. Test images are also located in IODP_Share > PhysProps > Testing and Calibration Notes > SHIL > SHIL New Lighting Test Images X384. The test images in the 'Images for Houpt' folder on scratch from testing in March were included in data1 folder. Tests are detailed below:

  • Test 1: F-Stop = f/16. Used the same calibration as X384. Sections were measured at 3 speeds: 9.4, 8.0, and 4.0

  • Test 2: F-Stop = f/16. The camera was re-calibrated, no physical settings were changed. Sections were measured at: 7.0, 4.0 (owing to new calibration speed must be lower than 8.0)

  • Test 3: F-Stop = f/16. The same calibration as test #2 but the gain was dropped down by 20, no physical settings were changed. Sections measured at: 7.0, 4.0

  • Test 4: F-stop = f/11. Camera re-calibrated. Sections measured at: 7.0, 4.0

  • Test 5: F-Stop = f/22. Camera re-calibrated. Sections measured at: 7.0, 4.0

  • Test 6: F-Stop = f/16. Lights moved up ~ 8 mm. Camera re-calibrated. Sections measured at: 7.0, 4.0

  • Test 7: F-Stop = f/16. Lights still ~8mm higher. Camera re-calibrated again. Sections measured at: 7.0, 4.0

  • Test 8: F-Stop = f/16. Lights moved down ~ 8mm from original position. Camera re-calibrated. Sections measured at: 7.0, 4.0

X387P/387T

• New LED lights were installed. IMS code was modified to accommodate this change.

• Copper heat dissipation bars were added to the camera.

• A frame was built by Jurie Kotze and Zenon Mateo to cover the shine of the lights, to protect users and to protect the equipment.

• Onboard cameras table was added to Confluence.

Technical Service (See ETs report)

• A fidget relay board was installed to control light switch with LabVIEW.

• An ON/OFF switch was install to control the lights manually.

X378/378S

• During 378 port call in Fiji vertical lines were noted in the SHIL images.  Bill Mills suggested to go into the JAI camera setup and re-run pixel flat and the other pixel correction (forgot name).  This is where you must de-focus the lens when shooting the grey card.  A full re-calibration was completed on Jan 8, 2020  and the issue was resolved.

X378T

• A new LED light source was tested during port call with good results. This new light source will need a circulating cooling system, much like that of the X-ray.

• 1x large 75 V capacitors in Power Supply box was replaced (ETs preventive maintenance).

• The fan on the Power Supply box was replaced for a new one (ETs preventive maintenance).

• By default the instrument is set for imaging the archive half and will not allow you to scan a working label. If you want to take a picture of a working half you need to go to the MANUAL tag and select W (working) into the Section Half label (See the picture bellow). Once you have selected the W (working), you will not be able to scan an archive half; in order to do that you need to go back into the MANUAL tag and re-select A (archive).  
  
  

X385

• Oct 2, 2019 Problem: Images were showing compressed lines at the very top of a section or the very end of a 150cm section.  Speed of down core motion was lowered and resolved issue sometimes.  Grinding sound could be heard when camera reached end of track.  Solution: One of the two bolts connecting the camera carriage to the linear encoder was loose causing them to bind at the end of the tracks.  The bolt was tightened.  The rod was removed from the encoder and an accumulation of dirt and dust was found inside the seals and cleaned.  This is likely a build-up of a few years since the rod isn't removed often. 

X379T/385T

• The SHIL station were dismantled in order to install new shelves that the ETs had made.

• Computers and instruments were disconnected and connected again at the end of the maintenance.

• Wiring was organized in a more practical, efficient and easy way to access for doing maintenance of the station.

• Routine maintenance on the NSK actuators were perform: clean and lubricate travel screw, inspect belts and tension correctly, check set screws on gears.

X383

• High Resolution RGB data was not being output.  A correction made to the code by removing the DECIMATE.vi from within the C:\IMS-10\Plug-ins\I_PI_aJAI-CV-L107CL\IMAGE RGB Graph.vi based on Bill Mill's advice corrected the issue.

X382

No Issues

X379

No Issues

X368T

No Issues

X378P/368X

• Nov 25, 2018 SHIL - default is to print VCD, should be off as default.

• SHIL - Had issues uploading, developers fixed this issue, see developer report

• SHIL- On the start of scanning a label, the behavior should be scan label and the program should pull a LIMS length, currently it is defaulting to the label length.

INSTRUMENT RESOURCES

Laboratory Manuals

• SHIL Quick Start Guide

• SHIL User Guide

• SHIL Light System

Vendor Manuals

• ADD ADDITIONAL INFO AND VENDOR MANUALS
  
  

Others

• SHIL Cameras Onboard List

• RGB Extraction Methods