Files changed (2) hide show
  1. README.md +62 -10
  2. cover.png +3 -0
README.md CHANGED
@@ -9,7 +9,7 @@ tags:
9
  - navigation
10
  - rosbag
11
  - terrasentia
12
- pretty_name: TerraSentia Under-Canopy Navigation
13
  configs:
14
  - config_name: cpp
15
  data_files: cpp.parquet
@@ -17,18 +17,73 @@ configs:
17
  data_files: crow.parquet
18
  - config_name: lecrop
19
  data_files: lecrop.parquet
 
20
  ---
21
 
22
- # TerraSentia Under-Canopy Navigation
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
 
24
  Navigation datasets from a **TerraSentia** agricultural robot driving
25
  **under the canopy**, extracted from ROS1 bags. Three sources/controllers:
26
 
27
  | Config | Episodes | Description |
28
  |---|---:|---|
29
- | `cpp` | 143 | Crop-follow / pure-pursuit (logs `path`) |
30
- | `crow` | 31 | iLQR controller (logs `crop_lines`, `goal`, `ilqr_time`) |
31
- | `lecrop` | 137 | MPPI/RL + vision (logs `dist_err`, `head_err`, `mppi_*`, `keypoint`) |
32
 
33
  ```python
34
  from datasets import load_dataset
@@ -93,9 +148,6 @@ Videos (`Video`, MP4/H.264, ~10 fps; `<g>_t` = per-frame time):
93
  > Source-specific columns are null when an episode does not have them (e.g. some
94
  > `cropfollowpp_lecropfollow` bags inside `cpp` carry `lecrop` columns).
95
 
96
- ## Provenance
97
 
98
- Built from the bag folders `cpp_runs`, `crow_runs`, `lecrop_runs` (rebuttal and
99
- named runs aggregated as one; exact filename duplicates were de-duplicated).
100
- Pipeline: `rosbags` → motion-based segmentation → per-episode video + signals →
101
- Parquet. Heavy raw sensors (depth / LiDAR point clouds) were excluded.
 
9
  - navigation
10
  - rosbag
11
  - terrasentia
12
+ pretty_name: LeCropFollow
13
  configs:
14
  - config_name: cpp
15
  data_files: cpp.parquet
 
17
  data_files: crow.parquet
18
  - config_name: lecrop
19
  data_files: lecrop.parquet
20
+
21
  ---
22
 
23
+ <h1 align="center">
24
+ LeCropFollow
25
+ </h1>
26
+
27
+ <h3 align="center">
28
+ Latent Space Planning for Navigation in Unstructured Crop Fields
29
+ </h3>
30
+
31
+ <p align="center">
32
+ <strong>Felipe Tommaselli</strong><sup>1</sup> &middot;
33
+ <strong>Francisco Affonso</strong><sup>2</sup> &middot;
34
+ <strong>Arthur Pompeu</strong><sup>1</sup> &middot;
35
+ <strong>Gianluca Capezzuto</strong><sup>1</sup><br>
36
+ <strong>Arun Narenthiran Sivakumar</strong><sup>2</sup> &middot;
37
+ <strong>Girish Chowdhary</strong><sup>2</sup> &middot;
38
+ <strong>Marcelo Becker</strong><sup>1</sup>
39
+ </p>
40
+
41
+ <p align="center">
42
+ <sup>1</sup> University of Sao Paulo &nbsp;&nbsp;
43
+ <sup>2</sup> University of Illinois Urbana-Champaign
44
+ </p>
45
+
46
+ <p align="center">
47
+ <em>IEEE Robotics and Automation Letters, 2026</em>
48
+ </p>
49
+
50
+ <p align="center">
51
+ <a href="#">
52
+ <img src="https://img.shields.io/badge/Paper-PDF-b31b1b?style=flat-square&logo=arxiv&logoColor=white" alt="Official Paper (soon)">
53
+ </a>&nbsp;
54
+ <a href="#">
55
+ <img src="https://img.shields.io/badge/arXiv-2026.XXXXX-b31b1b?style=flat-square&logo=arxiv&logoColor=white" alt="arXiv">
56
+ </a>&nbsp;
57
+ <a href="https://felipe-tommaselli.github.io/lecropfollow/">
58
+ <img src="https://img.shields.io/badge/Project-Page-4285F4?style=flat-square&logo=google-chrome&logoColor=white" alt="Project Page">
59
+ </a>&nbsp;
60
+ <a href="https://youtu.be/hV1fDjQsgOs">
61
+ <img src="https://img.shields.io/badge/Video-YouTube-FF0000?style=flat-square&logo=youtube&logoColor=white" alt="Video">
62
+ </a>&nbsp;
63
+ <a href="https://huggingface.co/datasets/arthurpompeu/lecrop-data">
64
+ <img src="https://img.shields.io/badge/Data-HuggingFace-FFD21E?style=flat-square&logo=huggingface&logoColor=white" alt="Data">
65
+ </a>&nbsp;
66
+ <a href="https://wandb.ai/arthurpompeu-usp-universidade-de-s-o-paulo/Cropfollow-input/reports/Lecropfollow-Report--VmlldzoxNzM3ODg4Mw">
67
+ <img src="https://img.shields.io/badge/Models-W%26B-FFBE00?style=flat-square&logo=weightsandbiases&logoColor=white" alt="Models">
68
+ </a>
69
+ </p>
70
+
71
+ ---
72
+
73
+ ## Models
74
+
75
+ Please check the `Files and versions` for our most up-to-date models. For more information, check: https://github.com/Felipe-Tommaselli/lecropfollow
76
+
77
+ ## Data
78
 
79
  Navigation datasets from a **TerraSentia** agricultural robot driving
80
  **under the canopy**, extracted from ROS1 bags. Three sources/controllers:
81
 
82
  | Config | Episodes | Description |
83
  |---|---:|---|
84
+ | CropFollow++ (`cpp`) | 143 | Crop-follow / pure-pursuit (logs `path`) |
85
+ | CROW (`crow`) | 31 | iLQR controller (logs `crop_lines`, `goal`, `ilqr_time`) |
86
+ | LeCropFollow (`lecrop`) | 137 | MPPI/RL + vision (logs `dist_err`, `head_err`, `mppi_*`, `keypoint`) |
87
 
88
  ```python
89
  from datasets import load_dataset
 
148
  > Source-specific columns are null when an episode does not have them (e.g. some
149
  > `cropfollowpp_lecropfollow` bags inside `cpp` carry `lecrop` columns).
150
 
151
+ ## License
152
 
153
+ Code is released under the [MIT License](LICENSE). The paper is published under CC BY.
 
 
 
cover.png ADDED

Git LFS Details

  • SHA256: 686673c65be43ea68ade1d8367e9898c279c039ca5410ed27398b7c7e4a27db3
  • Pointer size: 132 Bytes
  • Size of remote file: 4.39 MB