A webcam camera has been purchased for capturing IR pictures of the telescope and general observatory interior during operations.
This has been a long overdue purchase, and will provide much needed support for remote operation and monitoring of the Observatory.  Intention is that the camera will resolve ambiguities that might arise concerning the telescopes true position and to check and potentially identify issue regarding cable behaviour during operation and help diagnose other issues arising in the observatory.   The camera also provides a security function. 

Camera
The camera is :

ELP 1080P Night Vision Web Camera with IR LED, 2  Megapixel Full HD CMOS OV2710 Sensor UVC Webcam
compatible for Windows/Linux/Raspberry pi, 2MP Outdoor

ELP Product Page
Amazon Product page

Features

1. Use highly advanced 1/2.7 OV2710 CMOS Sensor
2. Low light performance , 0.05 Lux
3. Equiped with 1/2.7" 3.6mm 1080P high defination lens ,other lens optinal
4. High speed camera with high frame rate 640X480@120fps, 1280X720@60fps,1920X1080@30fps
5. Mainly used for advertising, video conferencing, high-definition monitoring ,Video doorbell, video intercom and other fields .
6. Support USB2.0 OTG protocal, can connect with OTG equipements
7. Full HD Night Vision Function, Excellent colour effect.
8. High compatibility , support Android, Linux, Windows, MAC OS
   
   See camera specifications below.

Arrival and Installation
The camera arrived 2021-01-15 and after indoor testing was loosely installed in observatory and connected to the powered USB Hub that is attached to the AllSky/Weather Computer. 

View Point
The Camera is currently loosely positioned so that it looks due south and towards the rear of the LX200 Main telescope (when parked ) with a view that is elevated to see as much of the telescope and dome interior as possible.  

The camera is designed to be mounted on a ceiling or on a high wall looking down,  but instead it is loosely mounted on a shelf and looks up. This means that raw frames from the camera are orientated upside down and have to be rotated 180 deg for normal display purposes.

The camera's viewing angle is 90 deg.   Whilst a slightly wider angle would have allowed a wider view of the observatory the 90 deg angle seems to be ok.   For this specific application it would have been handy if the camera had greater vertical viewing angle in order to see more of the dome north of zenith and see more of the pier.  A shame that the camera's FOV (aspect ratio 1920/1080 = 1.78) wasn't more square than it is.

Once optimal placement position is determined the camera may be fixed in a more permanent fashion.

Daytime view of the observatory as seen by the Camera (with IR)
Picture shown with a Gamma setting of 1.5.  Picture shows outside light filtering through the white fibreglass walls. No light filters through the dome roof  (which is painted matt black on inside), apart from some light evident around the margins of the shutter.
Telescope is in its normal park position (pointing due south with the OTA horizontal)

The viewing angle is 90 deg x 51 deg. The vertical view seen by the camera is illustrated on the observatory engineering cross-section.
The drawing and the camera picture both show that when the scope is parked, the view is almost face on to the fork arms and shows the bottom surface of the telescope (with counterweight bar & weights) and the bottom side of the imaging train/camera. are seen.

 

IR
The camera is provided with an IR Led Board. The IR lights are automatically always on when the camera is connected to USB Hub and it is dark (observatory closed without lights or nightime with observatory open at nighttime).   The power of IR lights is probably low to provide the camera's claimed night vision up to 10m, but are adequate for the 2.2m diameter observatory.

At the moment I haven't found a way of turning the IR lights off if/when required.  There is a concern that the IR lights might compromise light frames captured through the ST-10 camera on Main Scope or through the ASI ZWO 178MC camera on the Secondary Scope.    Initial indications suggest that this is not a problem, but detail tests need to be conducted in order to definitively confirm this.

Night-time raw pictures from the camera are too dark in the image's shadow areas, suggesting that image's gamma parameter need to be changed to bring out detail in dark areas.  Athough Brightness/Contrast/Color saturation /Definition/Gamma/WB are stated in the camera's specification to be adjustable parameters I haven't found a way to programatically adjust them yet.   Until this can be resolved images will instead be adjusted when after capturing them through bitmap adjustment.

NightTime view of the observatory as seen by the Camera (with IR)
Picture shown with a Gamma setting of 1.5. 
Telescope is in its normal park position (pointing due south with the OTA horizontal)Telescope is in park position.

Software
A custom built application AstroObsCam will be used to operate the camera.

Issues
Issues that have surfaced during first 6-9 months of operation :

• There can be IR leakage onto the Telescope Camera CCD Chip when the telescope is in certain positions (See Notes ObsCam - IR leakage onto CCD Images (tests and solution), 2021-04-15).  Whilst these were largely eliminated by wrapping a old jumper around the camera to block leak paths, there is still the odd position where leak occurs.
   
  Issue may be solved in future through the use of a Pegasus USB Control Box which will provide the ability to switch off power to the camera when the telescope is in particular positions where leakage is still able to reach the camera.
   
  
• The camera will sometimes lock/freeze and not provide fresh frames to the AstroObsCam program. After stopping and restarting AstroObsCam ta Windows 'Camera not found/not available' dialog oops up when trying  to reconnect to the camera. The only solution to this issue is to visit the obsevatory and then unattach/reattach the USB cable going to the Camera.  This is annoying as one may not have noticed that the ObsCam service had gone offline such that fresh ObsCam pictures aren't getting posted on the Observatory Status/Current Session web page and that the observatory has to be visited to resolve the problem. 
  A Pegasus USB Control Hub has been ordered 2021-10-13 in order to provide the observatory to be able to reset the camera when this problem occurs and looking further forward to give the observatory the tools to automatically restart the camera itself  (Note : issue added 2021-10-14)

Specification

Element	Description
Optional models	ELP-USBFHD01M-DL36
Sensor	1/2.7" CMOS OV2710
Lens size	200W, 2.0megapixel, 1920(H) x 1080(V) pixels
Max. Resolution	1920X1080
Picture format	MJPEG
USB Protocol	USB2.0 HS/FS, USB1.1 FS
Support free driver	USB Video Class(UVC)1.1
Auto exposure AEC	Support
Auto white balance	Support
Frame rate	1920 (H) x 1080 (V) pixels MJPEG 30fps YUY2 6fps 1280 (H) x 1024 (V) pixels MJPEG 30fps YUY2 6fps 1280 (H) x 720 (V) pixels MJPEG 60fps YUY2 9fps 1024 (H) x 768 (V) pixels MJPEG 30fps YUY2 9fps 800 (H) x 600 (V) pixels MJPEG 60fps YUY2 21fps 640 (H) x 480 (V) pixels MJPEG 120fps YUY2 30fps 352(H) x 288 (V) pixels MJPEG 120fps YUY2 30fps 320 (H) x 240 (V) pixels MJPEG 120fps YUY2 30fps
Adjustable parameters	Brightness/Contrast/Color saturation /Definition/Gamma/WB
Night vision	With IR LED Board, support night vision
View angle	Standard 3.6mm lens, FOV 90degree
Voltage	DC 5V
Current	150mA
Work temperature	-20 to +70 Degrees
USB Cable length	3m USB Cable
Support OS	WinXP/Vista/WIN7/WIN8 Linux with UVC Mac-OS X 10.4.8 or later Wince with UVC Android 4.0 or above

AstroObsCam is a new AstroSuite application for capturing IR webcam pictures of the telescope and general observatory interior during operations with associated facilities for uploading and display of pictures on current session web page.  Intention is that the camera's pictures will resolve ambiguities that might arise concerning the telescope's true position vs reported position, to check & potentially identify issues regarding cable behaviour during operations and help diagnose any other issues concerning the observatory. The software will also provide a security function. 

Software
The application is a VB.Net program based on a sample program by IronRazerz (on social.msdn.microsoft.com) program and makes use of the .NET c# directshow library file DirectShowLib-2005.dll from DirectShow.Net V2.1 (2010-02-20)  (from directshownet.sourceforge.net).  Prototype and the first release version of AstroObsCam was produced with 24 hours.  The first release version (1.0) and was installed on the AllSky/Weather Computer for live testing during session S859 (2021-02-16).

In additional to general monitoring and remote viewing of the AllSky/Weather Computer screen the application takes requests from AstroAllSky/AstroMain to Capture Images at specific points during operation.   AstroAllSky manages the regular upload of ObsCam.jpg pictures to the Astro-RichWeb website for convenient remote monitoring.

AstroMove has been updated to support easy backup of AstroObsCam software project,  transfer of new versions of AstroObsCam program to the AllSky/Weather Computer and the download of ObsCam pictures from the observatory for a given observing session.

WebCam
The application is setup to use 'HD USB Camera'  by default, in order to access a ELP WebCam camera (ELP 1080P Night Vision Web Camera with IR) that has been recently purchased (see above)

This camera is designed to be mounted on a ceiling or high wall looking down,  but instead it is loosely mounted on a shelf and looks up. This means that raw pictures from the camera are upside down and have to be rotated 180 deg for normal display..

During the first live testing of the WebCam/AstroObsCam it was found that night-time pictures from the camera are too dark in shadow and distant areas. This requirs that image's gamma parameter needs to be changed to bring out detail in dark/distant areas.  Although Brightness/Contrast/Color saturation /Definition/Gamma/WB are stated in the camera's specification to be adjustable parameters I haven't found a way to programatically adjust them.   Until this can be resolved images will instead be adjusted when either capturing them (through bitmap adjustment) or when saving them to file (through ImageMagicK).

Further Development (1)
Although the first live test of the AstroObsCam in session S859 (2021-02-16) was a success further improvements are needed and/or desired

• Lighten images. Specifically lighten the dark/distant areas of the image to bring out detail
• Annotate images. Annotate images with date/time and telescope target, operation and reported position.
• Catch GDI exceptions. JIT debugging of first live session caught  2-3 non-fatal exceptions that required user intervention to accept & continue.  Extra exception trapping is required to gracefully handle this exceptions
• Add Regular Captures. Add regular image capture 7 upload outside of session hours.
• Centre Line. Add checkbox option to draw a vertical centre line on to ObsCam Picture to facilitate camera position/orientation
• Compass directions. Add compass directions (East, South, West, Zenith) to ObsCam Picture to facilitate user orientation.
• Enhance Security Features.   Details are not discussed here.

Further Development (2)

• Logfile.  Add Logfile so that timing/logic issues can be identified/understood.
• AI Check.  Add AI to tell when the telescope appears to be in its Park position and when it is not
• Annotation Options.  Add option to show Annotation or not
  
  Traffic Lights show status of tasks as at 2021-02-24 (AstroObsCam 1.4.5)
  
  

A screen capture of the new application (v1.4.3, 2021-02-20) is shown below.

Application now uses a TCP/IP connection to the main Observatory Control Program (AstroMain) for accessing Session, Environment, Dome & Telescope information.  Annotation modified to use a sky blue colour instead of white and display a wider set of data. (this is ok at night but is too light during daytime when the observatory walls are paler during to filtering of sunlight from outside.

Link to latest ObsCam image.

Timing Stats
A number of operations were timed to see which operations take a particular long time and reduce the number of processed frames per second.

Rotate Full-Frame by 180 deg  : 18 ms
Apply Gamma : 370 ms    ** Little opportunity to reduce
Add Annotation : 1 ms

Create 1/10th Size Bitmap : 20ms  (same time including copy to Pic Box)
Motion Detection 400ms   ** Opportunity to reduce
Copy Half-Sized JpPicture to Secure Server Location : 500ms  ** Opportunity to offload to a separate thread

Current Issues

• Difference between nighttime and daytime pictures mislead the Telescope Position AI  : Brightness changes between daytime and nightime produce incorrect interpretations of whether the telescope is at its park position or not. When the picture is brighter (as it is during daytime) the scope is seen as "Not Parked" rather than "Parked".   Picture has to be normalised in some way.   This can't use the whole image as any changes in the scope position will affect the normalisation. Normalisation needs to be independent of the scope's position.
   
• Rapid fluctuation in daytime brightness associated with cloud passing across the sun cause False Intruder Alerts. Brightness changes between daytime and nightime produce incorrect AI results. When the picture is suddenly much brighter or much darker than the reference picture taken 2-3 s previously this is picked up as motion and triggers an intruder alert.  The affect of daytime brightness fluctuations may or may not be exaggerated by the use of a 1.5 gamma setting which is applied to each frame (for improving nightime detail) before Motion Detection is examined.
   
• ObsCam will find that process for capturing a new frame stops (locks up) / Camera becomes unavailable.  This seems to be a camera issue.  Closing and then restarting ObsCam, results in a Windows 'Camera not found/not available' dialog when trying  to reconnect to camera. The only solution to this is to visit the obsevatory and then unattach/reattach the USB cable going to the Camera.  This is annoying as one may not have noticed that the ObsCam service had gone offline such that fresh ObsCam pictures weren't getting posted on the Observatory Status/Current Session web page and that the observatory has to be visited to resolve the problem.  A Pegasus USB Control Hub has been ordered 2021-10-13 in order to provide the observatory to be able to reset the camera when this problem occurs and looking further forward to give the observatory the tools to automatically restart the camera itself  (Note : issue added 2021-10-14)