my scripts

A simple monocular visual odometry (part of vSLAM) by ORB keypoints with initialization, tracking, local map and bundle adjustment. (WARNING: Hi, I'm sorry that this project is tuned for course dem…

Run 3 scripts to (1) Synthesize images (by putting few template images onto backgrounds), (2) Train YOLOv3, and (3) Detect objects for: one image, images, video, webcam, or ROS topic.

Classification of 11 types of audio clips using MFCCs features and LSTM. Pretrained on Speech Command Dataset with intensive data augmentation.

Store some images, gifs, etc.

Implement: Sobel; Canny; Harris; Hough line; Fit line; RANSAC.

Apply ML to the skeletons from OpenPose; 9 actions; multiple people. (WARNING: I'm sorry that this is only good for course demo, not for real world applications !!! Those ary very difficult !!!)

2 Python API functions for point cloud conversion between Open3D and ROS. Compatible for XYZ and XYZRGB point type.

Visualize 3d humans' skeletons(body+hands) in ros rviz. The 2d joints are detected by openpose; The depth is from depth image.

A python script to publish color or depth images from a folder to ROS topic.

Python nodes to publish/subscribe RGB-D images and their point clouds (or any of them) to/from ROS topics.

Press key to record color/depth images from ROS topics or Realsense. Key `a` for saving single image; `s` for starting continuous recording; `d` for stop recording. `q` for quit.

Run one script and press 's'/'d' to save your laptop's camera images to disk. Two versions: (1) Python, and (2) ROS node.

ROS services for controlling Turtlebot3 to target pose by `Move to Pose` algorithm.

A python node to detect planes from depth image by using RANSAC algorithm. Input/Output from/to ROS topics.

(1) Press key to record audio; (2) Speak a word to microphone; (3) Finally, see the classification result on GUI and ROS topic.

Which object a person is pointing at? Detect it by using YOLO, Openpose and depth image (under customized scene).

Mathematica API for simulating the dynamics and collision of planar multi-link objects (by Euler-Lagrange equation).

Use a robot arm (Baxter) mounted with a depth camera to scan an object's 3D model.

A toy project of using voice to tell a Turtlebot Robot to detect and move to target, achieved by 4 components (1) speech classification, (2) object detection, (3) plane detection, and (4) control o…

Use voice to tell robot the target, then the robot detects it and moves there. (LSTM, YOLO, Plane detection, Motion planning, ...)

Use "optical flow", a simple "background subtraction" and cv2's "background subtraction" to get the mask of the moving objects

A toy project: Detect my hand grasping object in the video. Backbone algorithms: SiamMask, Mask_RCNN, OpenPose

Coding: ①Path planning: RRT*, A*; ② Tracking: Optimization, PurePursuit, FollowLine. ③Planning and control on a mobile manipulator

Put masked object onto background images randomly to generate images. Train Yolo3.

Only my VO project's test data and results

(1) 3D scan object by Baxter. (2) Label objects automatically by depth camera and (3) train Yolo. (4) [TODO; NOT DONE YET!!!] Finally, detect object and fit 3D model to know the 6D pose.

Put objects on a plane. Use depth camera to find them and add label (for training Yolo).