RCT node set-up » History » Version 11
Anchi Cheng, 09/05/2013 09:36 AM
| 1 | 1 | Amber Herold | h1. RCT node set-up |
|---|---|---|---|
| 2 | |||
| 3 | 4 | Anchi Cheng | h2. Acquisition Settings that is unique to this node: |
| 4 | 5 | Anchi Cheng | |
| 5 | 7 | Anchi Cheng | * Move Type: *Must be stage position or modeled stage position* |
| 6 | 4 | Anchi Cheng | * [[Iterative Stage Movement]] Option is ON |
| 7 | * Navigator Acceptable Tolerance should be set so that sufficient overlap between tilt pairs can be achieved. |
||
| 8 | 1 | Amber Herold | |
| 9 | h2. Testing feature finding algorithm |
||
| 10 | |||
| 11 | The default values in the RCT node may not work for your sample, trials should be made |
||
| 12 | to optimize the parameters. The good parameters should find hundreds, but not thousands of |
||
| 13 | features in your untilted image for good results. |
||
| 14 | |||
| 15 | |||
| 16 | 3 | Amber Herold | 1. Presets Manager> send the preset used in "Centered Square" to scope at the center |
| 17 | 1 | Amber Herold | of a square. |
| 18 | |||
| 19 | |||
| 20 | 3 | Amber Herold | 2. RCT/toolbar> Click on the "acquire" button at the far right to acquire a test image |
| 21 | 1 | Amber Herold | of which the feature finding algorithm will be applied after the acquisition. |
| 22 | |||
| 23 | |||
| 24 | 3 | Amber Herold | 3. RCT> Check the result of the numbers of feature found. A few hundred but not |
| 25 | 1 | Amber Herold | thousands would be a good settings. |
| 26 | |||
| 27 | |||
| 28 | 3 | Amber Herold | 4. RCT/Settings> Adjust Min Feature Size (most sensitive) and/or Max Feature Size or |
| 29 | 1 | Amber Herold | the filters and then repeat 2 and 3 to get better results. |
| 30 | |||
| 31 | 8 | Anchi Cheng | !good_feature_finding.png! !good_feature_finding_tilted.png! |
| 32 | Example of a good feature finding: There are good number of features. In addition, the obvious features, the center of most holes in the 3x3 hole patterns are identified. As a result, many of the features can be consistently identified after the distorting of tilting. |
||
| 33 | 1 | Amber Herold | |
| 34 | 8 | Anchi Cheng | !bad_feature_finding.png! !bad_feature_finding_tilted.png! |
| 35 | 9 | Anchi Cheng | Example of a bad feature finding: Even though the number of features are in the hundreds, the obvious features are not found because both Min and Max Feature Sizes are set too small. As a results, most features are lost after tilting (2nd panel) and unrelated features are found instead. |
| 36 | 1 | Amber Herold | |
| 37 | |||
| 38 | * Min/Max Feature Size: if < 1, it means faction of the image area. If > 1, |
||
| 39 | number of pixels. |
||
| 40 | |||
| 41 | |||
| 42 | * List of Tilts to Collect are in the order of data collection |
||
| 43 | |||
| 44 | |||
| 45 | * RCT node has its own drift monitoring after each tilt at the transformed focus |
||
| 46 | target position. Therefore, drift threshold and drift preset need to be defined. |
||
| 47 | 2 | Amber Herold | |
| 48 | 6 | Neil Voss | _see also [[RCT run protocol from a user]]_ |
| 49 | 2 | Amber Herold | |
| 50 | ______ |
||
| 51 | |||
| 52 | 11 | Anchi Cheng | [[Best grids for RCT|< Best grids for RCT]] | [[Best grids for RCT|Best grids for RCT >]] |
| 53 | 2 | Amber Herold | |
| 54 | ______ |