Releases: Team-Europa/Kibo-RPC_6th
Releases · Team-Europa/Kibo-RPC_6th
Final Version for Taiwan Preliminary
Contribution
- Fixed wrong estimating of Area 2 & 3 -- @milkteabox
- Fixed wrong estimating of Area 1 by adding offset and avoiding out-of-KIZ - @milkteabox and @Spig100
Note
HOPE WE CAN GET TO JAPAN!!!
First stable version for Preliminary
Contribution
- YOLOv11 implement -- @milkteabox
- Estimator and its offset -- @Spig100 @milkteabox
- Path testing -- @Spig100
- Batch testing helping -- @tyte103 @BioAppleLight
- Bugfix-- @milkteabox @Spig100
Known problem
This is the first version which can actually run stably. But still have Area 4 miss-scan issue in it. That feature is still testing.
v0.2.1: Bugfix and doc modification
This is the latest release in order to make our project easier to fix improve the path (which is the issue #3)
The model weight used in this release APK is 0524-deim_hgnetv2_x-K_Det-best_stg1.onnx
Bug Fixed
- Null pointer crash caused by
ItemDetectorUtils. getMaxFreqLandmarkItemData(issue #4) -- @milkteabox
Other modification
- Added hint for dev and simple team intro in
README.md-- @Spig100
v0.2.0: Realized DEIM-ONNX Detecting
Feature
- DEIM ONNX model runner and detecting function -- @milkteabox
- New path -- @milkteabox
TODO
- Improve the path to get more efficiently
- More accurate model with correct database (treasure and location combined)
v0.1.4 Full Rounding Completion Support
Feature Added / Modification
- Adding treasure item report -- @milkteabox
- Faster routine -- @milkteabox
- Adding rounding ended report feature -- @milkteabox
Note
This is the last version using YOLO model with NCNN interface for object detection. New pre-release v0.2.x would update to DEIM model with ONNX interface.
v0.1.3 Partial Mission Report
Feature Realized
- YOLO 11 NCNN detection & segmentation model implementation -- @milkteabox
- Reporting location item gotten -- @milkteabox
Issue Fixed
#2 -- @Spig100, @milkteabox
v0.1.2: Latest Version for dataset collection
Contribution
- Camera feature modification -- @milkteabox
- New test path -- @milkteabox
Known/Unsolved issue
issue #1
v0.1.1: Optimized Path
Feature
- Adding new testing path --@milkteabox
Note
This is the latest version for object detection data collecting.