InvertNet Newsletter, No. 1|
InvertDev: Updates from the Front Lines
From the Director (Chris Dietrich, InvertNet lead PI)
Greetings InvertNet Community! This is the first of our planned quarterly newsletters reporting on activities related to our efforts to digitize 55 million specimens from Upper Midwestern USA arthropod collections. Although our website has been operational for more than a year now and content is being added on almost a daily basis (943 resources and 126 registered users so far), this newsletter will be a periodic reminder that, yes, we're making steady progress. Our Phase-1 digitization workflows for slides and vials have been implemented and some collaborators have already captured images of a majority of their holdings in these two categories of objects. We're a bit behind schedule with the implementation of our workflow for digitizing whole drawers of pinned insects but prototype testing is underway. We need to make sure that we have all the kinks ironed out before we replicate the system for use at collaborating institutions. Many collaborators are anxious to get going on drawer digitization, but we're taking the time to get things right at the outset. Once our system has been fully implemented, digitization of drawers should move very quickly and we anticipate that even the largest collections will be able to capture images of all their drawers within a few person-months.
Meanwhile, most of our funded collaborators have been diligently curating and staging their collections in preparation for digitization and several have visited other collaborating institutions to help curate taxa for which they have expertise. These efforts will pay off in the long run because we are fixing many problems, such as misidentified/mislabeled specimens and unit trays, before the collections are digitized and exposed to the world.
In other news, we welcome the University of Hawaii (PI Dan Rubinoff), which recently received a grant from NSF to join InvertNet and digitize their insect collection. We hope to include many other institutions in the future and are building our cyberinfrastructure in such a way that it can be expanded to incorporate many more collections.
Drawer Digitization System
The prototype InvertNet Digitization system is operational! David Raila has been working constantly refining the hardware and software resulting in a system that is fast, easy to use, simple to ship, and able to be setup and operational quickly and easily. The system consists of a customized 3-PSS-PU parallel kinematic manipulator, more commonly referred to as a linear delta robot, a telecentric lens and camera, a LED lighting system, and software for automated precision control of camera positioning and image capture, and an integrated computer vision system that can detect and analyze images using a variety of features both during capture and after the final images are produced.
The system will automatically calibrate image positions using the computer vision software to detect the corners of specimen trays and using this input, calculate the optimal image positions to use during the capture. The system uses these positions to capture a grid of images across the tray, and includes automatic stitching capability to produce the final image. The images that you see below are of the prototype system in various states of refinement and in operation, and screen views of the machine control software, views of the live image from the capture camera of a single grid image, and some features of the vision system detecting image features and linear features from the live image. Please note that this is a working view of the prototype, intended to show the prototype machine, parts, wiring, and software at the time of the photos. The final system is intended to be delivered with motion and electrical parts pre-assembled, pre-wired, with software pre-loaded so that on-site assembly will consist of unboxing the pre-assembled parts, sliding the frame parts together and tightening the bolts, attaching the camera head, plugging in the modular cable connectors, and plugging it into the wall and turning the power on, from box to operation in 20 minutes. You may notice that the triangular format of the system is large; this is to accommodate both Cornell and USNM drawer dimensions. Currently, we are confirming the operation of a square format system which will be more compact and fit nicely on a desktop, and expect the performance to be equivalent or better than the triangular system.
We expect to finish testing and refining the drawer digitization system over the next month at the UI/INHS, after which we hope to begin setting up duplicate systems at collaborating institutions. As we anticipated, the total cost of the system, including a dedicated computer with pre-installed software, will be approximately $5000.
An example image from the camera/lens system, covering approximately 1.2"x1.4"
A view of the prototype machine positioning software showing a history of the paths that the machine has traveled, the camera position (white cone), and the grid program positions (blue). The InvertNet system will omit much of the complexity seen in the prototype interface.
A view of the computer vision system which is tracking object edges (black and white), and some image features (green markers on color image) in real-time.
The prototype Invertnet 3-PSS-PU PKM (delta robot) with exposed parts and wiring. The final system will include pre-wired and pre-assembled quick-connect components for operation in 20 minutes from opening the box.
Scan, Profile, Upload!
Vial and microscope slide image ingest pages and workflow demonstration videos are now available.
After months of grueling test runs, the ingest pages are online and accepting vial and microscope slide images! A few collaborators have made great progress towards unveiling their entire slide collections to the world with over 1000 slide tray images online!
We've filmed two workflow demonstration videos to streamline technician training and most importantly, ensure that standardized images are consistently produced by all collaborating institutions. The workflow videos and detailed descriptions of the procedures, equipment requirements and assembly are available in the resource section of the website.
Call for Content: Collaborator Buzz
Want to showcase your institution's progress along the InvertNet journey? Send photos and commentary to InvertNet.firstname.lastname@example.org for inclusion in the Collaborator Buzz column of future newsletters.Submission deadline for December edition is November 15th, 2012.
KSU's curatorial sweep through the Dakotas
Burke with collection manager, Rita Velez at SDSU (left), and Zolnerowich at NDSU (right)
Dr. Greg Zolnerowich and his team visited the collections at North Dakota State University and South Dakota State University in June. Zolnerowich and his graduate student, Alan Burke made strides in the Hymenoptera and Coleoptera holdings. At SDSU alone, they identified over 3000 specimens to family-level for a total of 60 families! Jan Metlevski directed his efforts towards organizing the Lepidoptera holdings of both museums.
Metlevski with Dr. Gerry Fauske at NDSU (left) and with Dr. Paul Johnson at SDSU (right)
"InvertNet: A new paradigm for digital access to insect collections"published as part of "No specimen left behind: mass digitization of natural history collections"ZooKeys Special issue, Chris Dietrich, et al., July 2012.
Dr. Jennifer Zaspel, new director of the Purdue Entomological Research Collection.
InvertNet TCN collaborator, formerly at the University of Wisconsin in Oshkosh, will begin her position as insect systematist (Lepidoptera: Erobidae) and director of the Purdue Entomological Research Collection.
Follow InvertNet on Twitter: @InvertNetTCN