Replies: 4 comments 6 replies
-
@Minims action: mqtt_vacuum_camera.reset_trims
target:
entity_id: camera.roger_camera
data: {} Reload Trims with this action will actually reset the trims and reload the camera at the same time. The point of this action is to be used when the vacuum actually rescan the floor. If the Floor - 1 isn't centered reset_trim might solve the image margining, unless if there are actually margins setup in the camera configuration.. that honestly if it is the case we might reset as well. Are you using the maploader on Rand256? If yes I would give it a try. When you use Integration > MQTT Vacuum Camera > Configure > Advanced options > Reset trims and then __reload __ the camera after the vacuum completed the scan or load a second map the image should be re-trim. Same as running the service. As I'm still running Rand256, and I know it is possible to save multiple maps... can you please be more precise and let me know if you intend to use the same vacuum on both floors? This is actually interesting and I don't know if we can "automate this process" using the Camera code. I will in any cases check (and change probably something on the reload "action") as per I noticed there could be some improvement, it looks that with more then one vacuum probably this action need to be target too. Hope this help, anytime, please let me know if we can improve something here. |
Beta Was this translation helpful? Give feedback.
-
... also.. I think.. I will change the code so that if there is a trims_reset and the vacuum is "cleaning" then the file with the new trim data should be saved at the end of the cycle Cleaning -> Docked = Save trims at the last image.. "snapshot time", instead of immediately. This should refresh the images "centring" them correctly. |
Beta Was this translation helpful? Give feedback.
-
Probably there is an issue with the "Action" reset_trims this is what I found by re-testing it.
the problem could be in the entity_id of the camera: can you please rename the Camera entity_id as per the robot name? in your case:
this should be equal to the name on the topic "vacuum_topic: valetudo/rockrobo" that by the way can be change in the vacuum configuration. So if your topic is "vacuum_topic: valetudo/roger" then definitely I need to probably open an issue here. |
Beta Was this translation helpful? Give feedback.
-
The "Action" reload camera will now reload all "mqtt_vacuum_camera" at once, and the "reset_trim" do not need the data anymore as will delete all auto_crop_*.json so that also if the camera name is different from the robot name the file will be deleted. I also implemented a delay to write again the file, the file will be written when the vacuum is docked, so.. if the vacuum is scanning a new map there is no trim locked, the image should (from what tested) be refresh correctly. I noticed from the test that the path and robot are draw but there is some minor adjustment on the segments to do as they do not draw during the re-scan (tested with the standard valetudo on my V1). From 2024.08.1 the Action "reset_trim" need to be format as following: action: mqtt_vacuum_camera.reset_trims
data: {} No more need to give the target. |
Beta Was this translation helpful? Give feedback.
-
Hello,
In 2024.8.0, does this action
is supposed to be the same as
Integration > MQTT Vacuum Camera > Configure > Advanced options > Reset trims ?
My problems is that i have configure camera for Floor-0. All is OK. Then i load the Floor-1 Map. The map is cropped and not centered.
I run : Integration > MQTT Vacuum Camera > Configure > Advanced options > Reset trims.
And the floor-1 map is OK.
I would expect the actions does the same thing, but it's currently not the case.
might be related to #207
but don't want to hijack this discussion.
Thanks for the help.
Beta Was this translation helpful? Give feedback.
All reactions