UCamCopy for UCCNC
User Manual — Version 2.400
User Manual — Version 2.400
UCamCopy is a vision-based probing and line-following plugin for the UCCNC machine-control software. It needs a webcam, preferably a “microscope” camera with fairly high magnification and adjustable focus. The camera image is processed with the EMGU/OpenCV libraries to find edges, lines, circles and enclosed shapes. Once an edge or target is found, UCamCopy drives the machine along it and records the path to a DXF file (and optionally a CSV file), so that you can reconstruct or copy existing geometry.
There is no scaling or dimensional calibration between the camera image and the machine. The image is used in a unit-less way to calculate position errors, and those errors drive the machine’s velocity through a closed-loop PID (proportional / integral / differential) controller using gains that you enter. The PID only generates velocity commands. Every dimension recorded in the DXF comes from UCCNC’s actual axis position in the active coordinate system, so the recorded data is as accurate as your machine’s steps-per-unit calibration.
UCamCopy shares its 3D viewer with the sibling plugin ProbeIt for UCCNC, so both plugins get the same viewer features.
UCamCopy relies on UCCNC to jog the machine. Jogging is disabled whenever UCCNC loses keyboard focus in Windows, so movements stop if you switch to another program while UCamCopy is running. There is no way around this other than not multi-tasking while video probing. You can move freely between UCCNC, the UCamCopy main window and the UCamCopy viewer window. UCamCopy notices when focus is lost and aborts the running operation properly.
UCamCopy tries to be as reliable as possible, but it operates your machine through UCCNC and its settings. Setting up UCCNC, your machine and the camera is your responsibility. You are responsible for checking and testing all signals, for the setup of your machine, and for becoming familiar with the software and its routines.
Machining is dangerous. Even though UCamCopy is only a video-probing utility and is not used while cutting, operate the machine as carefully as if you were machining. Never place your hands in the machine’s working envelope while video probing. Always turn off or disconnect the spindle motor power when a camera is mounted. Do not video probe at excessive feed rates, and use common sense. By using UCamCopy you accept all liability for damage or harm that could be caused by its use, to you and to your equipment. If you do not agree, do not use this software.
UCamCopy for UCCNC is still in development. It has been tested to make it as bug-free as possible, but it is not guaranteed to work with every setup. CraftyCNC is not liable for any damage or lost time resulting from its use. Probe-It, UCamCopy and CraftyCNC are in no way affiliated with CNCDrive or UCCNC.
The plugin works on both x86 and x64 computers running UCCNC, and is written for UCCNC versions that use .NET 4.0. UCCNC 2.2116 moved to a newer OpenTK graphics library, so there are two builds:
| UCamCopy version | Works with UCCNC |
|---|---|
| 2.2xx (for example 2.203) | 2.2115 and earlier |
| 2.3xx and 2.4xx (this manual) | 2.2116 and later |
If the wrong build is installed for your UCCNC version, UCamCopy will not run and will not appear in the list of plugins to enable. Install the one that matches your UCCNC.
You need a video camera connected and configured on the PC that runs UCCNC. Notes on the camera:
I have not tested many cameras; see the appendix for what I use and what I do not recommend. The plugin has been tested with UC100, UC300eth, UC400eth and AXBBe motion controllers, but any controller offered by CNCDrive should work equally well.
Install with the self-installer EXE file. The plugin is installed to the UCCNC plugins folder; its
settings, input and output folders live in plugins\UCamCopy\ (see appendix B).
UCamCopy installs in a demonstration mode. In Demo Mode all features are fully functional for a trial period, so that you can try it before buying. The window title bar and the About tab show the current status.
To get a license, send an email asking to purchase one, and include the License ID shown on the About tab. The license code is created and verified from that ID, which is generated from your UCCNC motion controller’s serial number. After payment, a passkey for that ID is emailed to you. Enter it in the Passkey box on the About tab and click Validate License Key. The status then shows Licensed and all limitations are lifted.
The Check Online for New Version button on the About tab compares your version with the latest one on the CraftyCNC website.
UCamCopy is a plugin and must be enabled in UCCNC before it can run.
If UCamCopy is not listed, something went wrong with the installation: check that UCamCopy.dll is in the
UCCNC plugins\ folder and that you installed the build that matches your UCCNC version (section 3.1).
To have UCamCopy appear when you ask for it, give it a button code that it listens for. In the plugins configuration window click Configure next to UCamCopy, type any number in the code box and click Set.
You can then assign that number in UCCNC’s I/O Trigger → HotKeys section, so that a key combination opens UCamCopy. In the example below the key code 110080 (Ctrl+Alt+P) triggers function number 3332, which opens UCamCopy because 3332 is the code entered in Figure 4.
You can also add a screen button in UCCNC and give it the same button code. The last way to open UCamCopy is to reopen the plugin configuration window and click Show.
UCamCopy is made of two main windows, plus a few settings windows:
Windows snap to one another when you drag them close together. When the viewer is stuck to the main window the pair moves as a group if you drag the main window; grab the viewer and pull to unstick it. UCamCopy works best at a screen resolution of 1280 × 1024 or higher.
The settings are not in tabs under the video any more. They are in separate windows that you open from the Settings menu, the first item of the menu strip above the camera view:
| Menu entry | Contents |
|---|---|
| Image Detection | Auto Tune, the HSV option, blur, Canny, circle and line-detection values (section 10) |
| Offset & Velocity | Camera/spindle offset moves and calibration (section 17), PID mode, gains and axis enables (section 13) |
| General Settings | Interface, safety and jogging, startup and updates, video overlay, detection colors, defaults and backup, DXF recording, live info (section 18) |
The windows have a blue frame, can be dragged anywhere (also onto a second monitor), and remember where you left them. They snap to the other UCamCopy windows too. Closing a window only hides it. The viewer has its own settings window, opened from its toolbar.
The side panels can be tucked away to give the video more room. The small arrow buttons on their edges (labelled Hide / Show) slide a panel away and bring it back. The viewer has the same strips for its two right-hand panels (Hide DRO / Info and Hide Layers / File Info). Which panels are shown, and how wide they are, is remembered.
The E-Stop button flashes while UCCNC is in E-Stop, so it is hard to miss.
The main window gives access to all of UCamCopy’s features, laid out in a way similar to UCCNC’s own screen:
The Special Operations menu holds the arc/circle collection dialog (section 12.3) and the G68 rotation setup. The Configuration Help menu opens two guides that walk you through the first setup: camera alignment, and calibrating axis velocity (section 13).
This is the shortest path from an unconfigured install to a traced part. Each step is described in detail in the later sections.
The left panel controls the camera and selects which stage of the image processing is displayed.
The Use HSV color for edges and detections checkbox (in the Image Detection window, next to Auto Tune) makes UCamCopy use color information instead of grayscale to find edges. This helps when your part looks similar to its background in grayscale. It is always better to create a color or shade difference yourself, for example by placing the part on light or dark paper.
I usually zoom in so that about 1 cm (0.5 inch) spans the width of the camera view. Even with 640 × 480 pixels that gives less than 0.001 inch (0.015 mm) per pixel, which makes UCamCopy accurate without any calibration.
EMGU/OpenCV captures the video and runs several steps to find lines, edges and circles:
Grayscale flow (HSV off): Capture → Grayscale → Blur → Mask (narrow region of the image) → Canny filter (edge detection) → Detections → output to drive the machine.
HSV color flow (HSV on): Capture → HSV conversion → Blur → Mask → Canny filter → Detections → output to drive the machine.
The Select View options let you look at each stage, so you can set thresholds more easily. During normal operation you will leave Composite Output selected.
The pictures below all use the same example part so you can compare the stages.
The original, unmodified camera image. Use it to check lighting and focus.
A black and white picture of the mask. Image processing is limited to what falls in the white area.
The most important view. It shows the edges the program finds after processing the image inside the masked region (yellow outline). Use the Edge (Canny) Threshold values and Thicken Edge (Canny) Lines to filter the edges until you get a strong, clean result like this one.
The grayscale image with blur applied. Available only when HSV is not checked. Adjust the Blur Amount in the Image Detection window to remove noise: too much blur makes edge detection difficult, too little leaves a lot of noise.
The original image converted to HSV color space and blurred. This can separate the image much better than grayscale, but needs extra thresholds to isolate the area of interest.
The HSV image needs the Color Detection Thresholding sliders adjusted. On the left below there are no thresholds and the areas are not separated well enough to detect an edge; on the right the thresholds give a clear edge.
The final image, with overlays of the DROs and cross-hairs (if turned on) and the detections that drive the machine. The dark blue arrow is the direction the machine moves when Run is on. Red points and light-blue lines are the detected edges and the endpoints used to calculate the movement.
Shows the four most useful views at once, so you can adjust thresholds without switching back and forth. With HSV on you see the HSV image, HSV threshold, masked Canny edges and composite output; with HSV off you see the gray image, gray threshold, masked Canny edges and composite output.
Open Settings → Image Detection. The window opens in a Simple view that shows only what most people need: Auto Tune and its Undo button, the Use HSV color checkbox, the blur value, the Canny low/high thresholds and, when HSV is on, the HSV color range sliders. Tick Advanced settings to show every detection setting.
| Setting | What it does |
|---|---|
| Blur Amount | Removes noise before edges are found. |
| Edge (Canny) Threshold Low / High | Limits which edges are kept. Raise them to ignore weak edges. |
| Thicken Edge (Canny) Lines | Makes detected edges thicker. Advanced. |
| Circle Threshold | How strong a circle must be to be detected by the circle detector. Advanced. |
| HSV color range | With HSV on, the range of hue, saturation and value that counts as the target. |
| Use custom line detection settings | When ticked, the Hough and merge values below are used. When unticked, built-in defaults are used. Auto Tune turns it on. Advanced. |
The line-detection values (Advanced view) control the line detectors and the line-merging algorithm:
Finding good blur, Canny and line-detection values by hand can take a while. Auto Tune sets them for you with one click.
Auto Tune needs the camera running (it asks you to start it first). You can still adjust any value by hand afterwards.
The mask limits the part of the image that is used for detection. It has two shapes (round and rectangular) and an adjustable size. When you edge-trace, the size and shape of the mask are important: they reject noise and keep the tracing on course. The mask follows the angle of the edge it finds, which helps with steering. Items inside the white area of the mask are “seen”; items in the black area are rejected.
When edge tracing, your goal is to adjust the image and the mask until you have essentially a single line with a target dot at each end inside the mask. That is the line that will be followed.
The detector buttons in the bottom panel choose what UCamCopy looks for:
The detector you will use most, for auto-tracing the outside edges of objects. It needs a clear edge in the image and a mask that limits detection to the area of interest.
You choose the direction the machine travels around the perimeter by double-clicking the image to re-orient the mask in the opposite direction:
You may have to help by positioning the part so the edge is initially centered, and by orienting the mask. In the next picture the edge is in the right place but is not detected, because the mask is not oriented over the edge:
When you are calibrated and ready to trace: switch to line mode, jog until the cross-hair is over an edge, and adjust the image and the mask until you have a clean line with target dots at both ends. The mask aligns itself with the edge and can go either way around the part (clockwise or counter-clockwise makes no difference). Double-clicking near the center of the image re-orients the mask direction.
You can start and stop the velocity and move around as needed to cover all the edges. If the line is lost or the machine gets off track, stop the velocity, jog back into place, adjust the video and mask, and enable velocity control again.
When enabled, the detector looks for the two most prominent lines in the image and uses their intersection as the target. Turn the mask off.
There are two ways to deal with holes and circles.
If the hole is small and fits completely in the camera view, select the circle detector and the machine drives to its center point. It only finds a location; the size of the circle cannot be determined from the camera image alone. Adjust the Circle Threshold until you get a reasonable detection (it may bounce around quite a bit). In the picture, the red dot is the center target and the dark blue arrow is where the machine will drive.
For geometry that is too big for the camera view, open Special Operations → Arc or Circle From Points. The dialog helps you collect circles and arcs into the DXF or find sizes and center points.
Follow the prompts: jog to at least three locations on the arc or circle and click Sample Point at each. With three points the center and the diameter/radius are calculated and shown. You can collect more points for a least-squares fit, and then use the result to:
For an arc you can only use three points: sample one end, a point along the arc, and then the opposite end.
Similar to the circle detector, but it finds shapes that are not perfect circles more easily. It is the detector I recommend when calibrating, because it is more reliable and stable than the circle detector and has no threshold to adjust.
Before going far with UCamCopy, set up the PID controllers (proportional / integral / differential) that drive the axes together with the velocity you choose. They are on the Offset & Velocity settings window (Settings → Offset & Velocity).
Under the Configuration Help menu, camera alignment and Calibrate Axis Velocity open guides that walk you through the initial setup.
Key points:
| Mode | How it works | Use it when |
|---|---|---|
| PID Only | One PID per axis, aimed at a target held a fixed distance from the current position. The X and Y PIDs calculate separate velocity commands from the difference between target and current position. | The part has sharp corners or lots of twists and turns. |
| PID w/ Centering PID | As above, plus two “Center PID” loops that reduce the following error between the closest point of the edge and the current position. Recommended. Can be unstable if the center PID is poorly tuned or aggressive; set its max output to 5 or 10% (the limit is 50%). | Most work. |
| Angle/Velocity | Uses the Center PID loops to put the closest point of the edge at the target, and takes velocity out of PID control: speed and direction come only from the slope of the detected edge. A simplified algorithm that may work better on some machines. | The part has long straight sides and few twists. |
PID is a classic negative-feedback control (servos, thermostats and so on). In UCamCopy it looks at the error in the camera image — how far the center of the picture is from a defined target — and sends velocity commands to UCCNC to drive that error to zero. kP, kI and kD are gain coefficients. They must be set before you let UCamCopy control your machine.
The total gain fed to the machine is the combination of the maximum velocity you enter and the PID gains for the normal and center PIDs. If gains are too high the machine oscillates; if too low, it is slow and the following error grows, which turns into measurement error.
Place a small dot on a piece of paper as a target, then:
When it works, the target dot is brought to the center of the view and stops without oscillating. You can then change the values, jog off center and run again to test them. You want the machine to snap quickly to the dot without oscillating or overshooting.
The PID Update Rate (in ms) is how often the PID loops are calculated and velocity commands are sent to UCCNC. I suggest about 100 ms (10 times per second). It should be slower than the camera’s frame rate by about 50%. For example, a camera running at 15 fps delivers a frame every 67 ms (1/15 = 0.067), so set the update time larger than that; 100 ms gives 10 updates per second. If your camera has a consistently high frame rate you can decrease the value: 50 ms (20 Hz) works with a 30 fps camera.
You can see the real frame rate in the status messages at the bottom of the window:
UCamCopy does not run routines step by step like a touch probe. Instead you use the Run/Stop button to give UCamCopy control of the axes or to take it away.
If there are no targets, jog the machine by hand to bring one into view. If the target is in view but not detected, adjust the image and the detections until it is.
Machine moves can be aborted, and fail safe: if UCamCopy loses focus or cannot move, it stops rather than guessing.
Once you have opened a file to record to, you can start and stop recording as often as you like. For example, you can record some perimeter feature (click Abort to stop), record a bore or edge feature elsewhere, set a new Z location, go back and record another perimeter, all to the same file. When you are done, save it all with Save DXF.
The top panel controls recording of DXF or CSV data, opening, saving and closing files, choosing the DXF layer to record on, and so on.
You can also record points manually and choose whether each point is joined to the previous one by a line (line join / break).
The Point Sample Distance is used during auto edge tracing and sets how far apart the recorded points are, which controls the resolution of the edge: a smaller number means finer resolution. Edges are recorded as points joined by straight lines. You can fit this data to polyline curves in your CAD or CAM software afterwards.
UCamCopy records to a CSV (comma-separated values) file that Excel can import, to a DXF file, or to both at once. Files are stored in the output folder of the
UCamCopy folder (for example C:\UCCNC\plugins\UCamCopy\output\).
Turn on the CSV checkbox, the DXF checkbox, or both. You are prompted for a file name; if you select both, the name is used for both formats. Once the DXF or CSV button shows ON and a name has been entered, you only have to trace the part: all data is recorded until you close the file. Starting a DXF recording creates (or selects) the DXF layer for you; use the Active Layer selector to record to a different layer.
If no DXF data has been recorded, Open opens an existing DXF file. You can open several files, which are appended together; new data you collect is merged into them. When you save, you can overwrite or save under a new name. As soon as a file is opened or data is recorded, the button works for saving until you clear the DXF buffer or save the file.
While a recording option is on, data keeps being collected and appended internally. After you finish probing a part, click Close DXF to close the file (you are asked to save if you have not). To throw away data you do not want, click Empty DXF, which clears it without saving.
You can click Save DXF at any time to make sure the results are on disk. If there is no file name yet you are prompted for one; if CSV and perimeter files are selected, they are saved with the same name as the DXF. Once the file has a name it is displayed, so you can confirm where the data is going.
To save the DXF data under a different name, use Save As on the viewer toolbar (or Save As… in the viewer’s right-click menu). Unlike Save DXF it writes the file whether or not the DXF button shows ON, and the new file becomes the current file, so a later Save DXF goes there (section 16.9).
The viewer shows the point cloud and DXF data measured and recorded while you probe, and the camera position in real time. It also shows information about the DXF, lets you control the layers and colors and choose the active layer to record on, and is a basic DXF editor: you can pick, delete, move to another layer and draw lines. Colors and visibility are set by you.
Each of the two right-hand panels has a vertical Hide / Show strip beside it. Click it and the panel slides away to give the 3D view more room; click again to bring it back. Drag the left edge of a panel to make it wider or narrower. Which panels are shown, and how wide, is remembered the next time the viewer opens.
A short right-click that does not drag opens the right-click menu instead of rotating.
Click a line, circle, arc or polyline to select it. It is drawn in the Selection Color (cyan by default) and its layer and coordinates are shown in the status bar. Before you click, the entity under the pointer is drawn in the Hover Color (orange by default). Open shapes (lines, arcs, open polylines) show a green Start dot and a red End dot, and an arc also shows a blue Center dot. Closed shapes such as circles have none of these.
While Select or Zoom Box is on, dragging draws the box instead of panning the view.
With entities selected, click Delete (or press Delete) to remove them from the DXF data. Undo (Ctrl+Z) takes back the last change and Redo (Ctrl+Y or Ctrl+Shift+Z) puts back what you just undid. The history covers deleting entities, drawing lines and moving entities to another layer, one change at a time. A new change clears the Redo history, and opening or clearing a DXF clears both.
Any edit marks the DXF as having unsaved changes, so the Save button on the main window becomes available.
Draw Line adds new lines to the DXF by clicking in the view. Click Draw Line on the toolbar (the button stays highlighted while the tool is on), then:
Snapping: when the pointer is within about 12 pixels of the start or end of an existing line, an orange dot appears on that endpoint and the status bar names it. The point you click is then exactly that endpoint, so lines join cleanly. Away from an endpoint the point is where you click, at Z = 0. Only the endpoints of lines are snap targets; arcs, circles and polylines are not.
New lines go on the active layer (shown in the Layers panel; one is created if there is none). Each line is one Undo step. Pan, rotate and zoom still work while the tool is on. Use the Top view for drawing: in an edge-on view such as Front or Right a click cannot be placed on the Z = 0 plane.
The layer list has these columns: Vis (click to show or hide the layer), Col (double-click to change its color), Del (delete the layer and its entities, after a confirmation), Layer Name (edit it to rename the layer) and Cnt (the number of entities on it). Add Layer creates a new layer and Refresh Layer List rebuilds the list. The Active Layer is the one that newly recorded data and newly drawn lines are put on. Layer names must be unique (upper and lower case count as the same): a name that is already in use is refused, both for a new layer and when renaming.
To move entities to another layer, select them, right-click one of them, choose Move to layer and pick the destination layer. The entities keep their positions; only their layer changes. They stay selected so you can see the result, and Undo moves them back. If the layer you want does not exist yet, create it first with Add Layer or New Layer… in the right-click menu.
To make a new layer, click Add Layer (or choose New Layer… in the right-click menu). UCamCopy asks for a name, then for a color. If you cancel the color dialog the layer is still created.
Moving entities, step by step: first select them (here part of the red Tri layer is selected, shown in cyan), then right-click one of them and open Move to layer, which lists every layer. Pick the destination.
Right-click (without dragging) in the 3D view to open a menu. If you right-click an entity that is not selected, it becomes the selection; if it is already part of the selection the whole selection is kept, so you can act on a group. The menu has:
Entries that do not apply, such as Delete with nothing selected, are greyed out.
The Display Options menu turns the Axis / WCS indicator, the grid and the current-position marker (Show Cur. Position) on and off. The colors and sizes are in the viewer’s settings window, opened from the toolbar. Its groups are collapsible; Viewer Colors and Sizes are open by default.
Click a color swatch to change that color. The settings are saved with the viewer settings when you click Save Settings on the main window.
Click Save As on the viewer toolbar, or Save As… in the right-click menu, to save the DXF data to a new file. The save dialog opens in the folder of the current file with its name filled in, ready to change. The data is written whether or not the DXF button on the main window shows ON, and the new file becomes the current file: the name shown on the main window changes to it, and a later Save DXF goes there. The original file is not changed. Save As writes only the DXF; the CSV and perimeter files are written by Save DXF.
You will most likely have the camera side-mounted to your spindle. The offset controls (in Settings → Offset & Velocity) let you set the offset between the spindle and the camera and then toggle the machine from one position to the other.
This is a good setup tool: use the camera and cross-hairs (or the detectors, or the arc/circle routines) to locate the origin you want, then move the spindle to that location and zero the DROs.
Each session, the first time you zero an axis or use a button to move between the spindle and the camera, you are asked which position you are currently using (camera or spindle). From then on the buttons only toggle between the two.
Settings → General Settings holds the options that are not about image detection. They are in collapsible groups in two columns; whether each group is open is saved with your other settings.
| Group | What is in it |
|---|---|
| Interface | Enable Tool Tips?, Enable Error Message/hints?, Group Minimize/Restore All Windows?, and the Config ‘Open’ Button Code button, which opens the dialog described in section 5.2. |
| Safety & Jogging | Disable Jogging in UCCNC interface? turns off jogging from UCCNC’s screen, the arrow keys (if assigned) and the fly-out jog panel, so UCamCopy is in control and a stray key press cannot jog the machine. Check for softlimits, abort if hit stops and reports if a soft limit is hit during a move. |
| Startup & Updates | Auto Save ‘settings’ on close? saves your settings when the plugin window is closed. Auto Check for New Version Enabled? checks the website for a newer version when the plugin starts (if you have a network connection). The Check Online for New Version button on the About tab works whether or not this is ticked. |
| Video Overlay | DROs on the image overlay, DRO color, cross-hair overlay, cross-hair color, line width, circle radius and number of cross-hair circles. |
| Detection Colors | The colors of the detected line, the target point and the move vector. |
| Defaults & Backup | Set Defaults – INCH, Set Defaults – MM, Save current as defaults… and Restore from backup…. |
| DXF Recording | Record Z on zero plane, the Z value to record as, circle size to draw on “record pt”, and cross-hairs on circle. |
| Live Info | The live detection information that is also shown at the bottom left of the main window: video info (image size, actual and processing frame rate), mask info (the mask angle) and line info (lines found, and lines left after merging similar ones). |
Click Save Settings on the main window to save your settings (the controls of the main window and the settings windows, and the viewer settings). Settings are only saved when you click the button, plus an automatic save when the plugin closes if Auto Save is on and the settings loaded properly. If the settings could not be read when the plugin started, nothing is auto-saved over them.
The first time UCamCopy runs with no saved settings, it asks whether you work in inch or metric and loads matching starting values. You can also choose a settings file instead. The starting values are sensible first-pass settings for both the main window and the viewer; the metric set has the feedrates and lengths converted (for example the probing velocity is 102 mm/min instead of 4 in/min, and the viewer’s grid is 127 mm across instead of 5 in). Please check the values against your machine, camera and zoom, especially the PID gains (section 13).
Everything is in plugins\UCamCopy\Settings\ inside your UCCNC folder:
| File / folder | What it is |
|---|---|
UCamCopy.xml, UCamCopyViewer.xml | The main window and viewer settings: every checkbox, number, text, combo, color and group state. |
UCamCopy.settings.xml | Window positions, sizes and panel state. On the first run, values from an older install are imported once. |
Backups\ | Rotating backups (the newest 10 of each file), made whenever a settings file is about to change. |
SettingsLoad.log | One line for every save, load and restore, plus anything that could not be applied and why. Look here first when “a setting didn’t stick”. |
WindowPositions.xml | Where each settings window was left. |
Defaults\ | The defaults you saved. |
Settings are written safely (a new file is written and checked before it replaces the old one), and a settings file that cannot be read is never overwritten by the defaults.
Restore from backup… (General Settings, Defaults & Backup) lists the backups and puts the one you choose back.
| Symptom | Check |
|---|---|
| UCamCopy is not in the plugin list | Is UCamCopy.dll in the UCCNC plugins folder? Is it the build that matches your UCCNC version (2.2xx for 2.2115 and earlier, 2.3xx/2.4xx for 2.2116 and later)? |
| The camera is not in the list | Plug it in, then click the refresh button next to the camera list. Make sure no other program has the camera open. |
| Run is on but the machine does not move | Moves only happen when a target is detected. Check that the overlay shows target points and a blue move vector; if not, adjust the image, thresholds and mask, or run Auto Tune. Check that UCCNC has keyboard focus and is not in E-Stop. |
| The machine goes the wrong way | Check the camera orientation (+X to the right, +Y up) and the axis definitions in UCCNC. See section 13. |
| The machine oscillates | The gains are too high. Reduce kP (and kD); see the tuning procedure in section 13.3. |
| Movement stops when I use another program | UCCNC jogging stops when UCCNC loses focus. Do not multi-task while running (section 2.1). |
| Edge is found but not traced | The mask is probably not oriented over the edge. Double-click the image to re-orient it. Check the mask size and Max Angle Delta. |
| Traced points are not in the DXF | Make sure the DXF button shows DXF ON and a file name has been entered (section 15.1). |
| Low frame rate | Check lighting: many cameras drop to about 10 fps in poor light. Compare Act FPS with Process FPS (section 13.4). |
| A setting was not remembered | Click Save Settings. Then check Settings\SettingsLoad.log, and use Restore from backup… if needed. |
| The program is not licensed after I entered the passkey | The passkey must match the License ID shown on the About tab, which is created from your motion controller’s serial number. |
| Action | How |
|---|---|
| Pan | Left mouse button + drag |
| Rotate | Right mouse button + drag |
| Zoom | Mouse wheel |
| Reset view | R (pointer over the 3D view) |
| Select an entity | Click; Ctrl/Shift-click to add or remove |
| Box select / zoom box | Select or Zoom Box tool, then drag; Alt removes |
| Clear selection / leave Draw Line | Esc |
| Delete | Delete |
| Undo / Redo | Ctrl+Z / Ctrl+Y (or Ctrl+Shift+Z) |
| Run / Stop (main window) | Space bar |
| Re-orient the mask direction (main window) | Double-click the image |
| Set a first HSV range (main window) | Shift-click the image |
| Path (inside the UCCNC folder) | Purpose |
|---|---|
Plugins\UCamCopy.dll | The plugin |
Plugins\UCamCopy\Settings\ | Settings, backups, defaults and log (section 18.4) |
Plugins\UCamCopy\output\ | Recorded DXF and CSV files |
Plugins\UCamCopy\input\ | Input files |
Plugins\UCamCopy\Images\ | Interface icons |
Plugins\UCamCopy\Manual\ | This manual (opened by the Help button) |
Plugins\UCamCopy\README.txt, UCAM_EULA.rtf | The install notes and the license agreement |
Emgu.CV.UI.dll, Emgu.CV.World.dll, ZedGraph.dll, DirectShowLib-2005.dll, MagnetWindows.dll, x64\, x86\ | Video and graphics support files, installed in the UCCNC folder itself and shared with ProbeIt |
For UCCNC 2.2116 and later. A broad round of stability fixes, viewer features and user-interface clean-up on top of V2.301.
V2.203 is for UCCNC 2.2115 and earlier; V2.301 is for 2.2116 and later. Only a minor bugfix (the license text box on the About tab was hidden); no other functional changes.
A dual release. UCCNC 2.2116 moved to a newer OpenTK graphics library (3.3.2) that made UCamCopy incompatible unless it referenced the newer library, so the 2.2xx versions are for UCCNC 2.2115 and earlier and the 2.3xx versions are for 2.2116 and later. The viewer was updated to match ProbeIt’s pan/zoom/rotate and gained snap-to-view.
Installer fix only (magnetwindows.dll was missing).
Added the 3D viewer, the DROs tab, opening existing DXFs and appending to them, CSV perimeter recording and default XML files for inch and metric; fixed the PID calculation that saturated both axes at 100% on diagonals; masks may be longer than 100 pixels.
V1.001 fixed the velocity input limit, added Shift-click to pick an HSV color, tool tips and a setting to turn them off, and removed auto-save on close because it could overwrite settings with empty data if UCCNC closed first. V1.000 was the initial demo release.
Cameras. I cannot make a direct recommendation on brand or model since I have only tried a few. In general, look for one that advertises 30 fps and has adjustable focus done with a knob at the back end.
What camera do I use? The one below, bought on Amazon. It is also sold by Oasis Scientific as model V1-PS.
Things I like:
Things I do not like: some background noise in the image, but I think that is common with all cameras of this type (CMOS sensors).
Any camera I absolutely do not recommend? Yes: the type with a focus wheel near the lens instead of at the back. The lens has play, so the image moves around and the focus does not stay adjusted. It is fine as a real microscope but does not work well mounted on a CNC. This type also had a lower frame rate and a noisier image, and a maximum resolution of only 1920 × 1080.
UCamCopy! Copyright 2023–2026 by Crafty CNC / Eric Brust. www.craftycnc.com