README.md

Robotic-Conductivity-Probe-Optimization

Overview

This toolkit provides robust methods for image segmentation and path planning, employing Meta AI's Segment Anything model and optimization techniques for efficient pathfinding.

Features

• Image segmentation using Meta AI's Segment Anything model.
• Watershed segmentation to refine image segmentation.
• Path planning with optimized probe poses.
• Evaluation of probe contact points within segmented regions.
• Maximizing spatial and angular variation of poses.

Dependencies

• numpy
• pandas
• matplotlib
• scipy
• opencv-python (cv2)
• scikit-learn
• segment_anything (Meta AI's Segment Anything Model)

Installation

Ensure all dependencies are installed using pip:

pip install numpy pandas matplotlib scipy opencv-python scikit-learn

Meta AI's segment_anything model is used to segment the pixels of each material.

• segment_anything must be installed separately as per its documentation.
• To run segment_anything, a checkpoint of model weights must be downloaded and its path must be specified to run Pose_and_Path_Optimization.ipynb.

Usage

Image Segmentation

generate_segments(image, checkpoint, model_type, min_size=None, max_size=None)

• image: Input image for segmentation.
• checkpoint: Model weights for Segment Anything Model.
• model_type: Type of the Segment Anything Model.
• min_size, max_size: Optional size constraints for segmentation.

Probe Pose Optimization

probe_contact(midpoint, rotation, probe_stroke_px)
generate_valid_poses(droplet, num_poses, max_angle, probe_stroke_px)
reward_function(droplet, poses, max_poses, max_angle, probe_stroke_px, verbose=False)

• These functions generate the optimal poses for a given material droplet.

Path Planning Optimization

path_planning(poses, noise_level, start=[0,0], optimization_rounds=1000)

• Path plans are optimized using a stochastic nearest neighbors approach at varying noise levels.

Examples

Files	Description
Camera_Intrinsics_Calibration.ipynb	A python notebook with an example of performing camera calibration.
Pose_and_Path_Optimization.ipynb	A python notebook with an example data set demonstrating optimization of poses and path plans.
robot_functions.py	A python file with all necessary functions to perform pose and path optimization.