Prepare a corkboard (the tape in the picture is A4 paper for positioning).
Place the P340 robotic arm in the center of the bottom side of the corkboard.
Reference distance between pen tip and pen clip: 22mm.
Build a camera stand and place the camera at the top to create a top-down view. The camera should be 330mm away from the wooden board. After powering on the robotic arm, connect the data cable and camera cable to the user's computer.
Notes: Python version must be 3.8 or higher.
pip install -r requirements.txtAPI Key Acquisition: Users need to create their own Tongyi Qianwen and Deepseek accounts and configure their own API keys. API key usage is charged; users should check their Tongyi Qianwen and Deepseek account balances before use.
Tongyi Qianwen API Key Application Reference Video: https://www.bilibili.com/video/BV1iHVazvEvS/?spm_id_from=333.337.search-card.all.click&vd_source=672e3f7240eaaca210b45e7c033dc45f
Deepseek API Key Application Reference Video**: https://www.bilibili.com/video/BV1vCcge7E6W/?spm_id_from=333.337.search-card.all.click&vd_source=672e3f7240eaaca210b45e7c033dc45f
Modify the config.yaml file in the /config directory to configure the software. Please ensure correct configuration before running the program. Enter your own Deepseek API key in the deepseek API key field, and enter your own Tongyi Qianwen API key in the qwen and qwen_vl API key fields. Enter the camera ID in the camera field. If using a laptop, the ID defaults to 1; if using a desktop computer, the ID is generally 0. Enter the actual serial port number of the robotic arm in the com_port field of the robot configuration file.
After ensuring the user's computer is connected to the network, run the main.py file, wait for the robotic arm to return to zero, and then tilt to one side.
Fix the answer sheet in the center of the corkboard, with the bottom of the answer sheet firmly against the robotic arm.
Enter 1 or 2 in the terminal and press Enter to confirm the mode. If you choose direct writing, enter the desired text in the terminal and press Enter. The robotic arm will then begin writing the user's input. If it's in AI answer mode, it will automatically open the camera to capture the content.
Then manually adjust the camera and corkboard so that the image capture interface is as shown in the image below (leaving approximately one-sixth of the image blank). Be sure to adjust according to this image to ensure the camera view is as consistent as possible with the reference image below; otherwise, the writing area will be incorrect.
Press the spacebar to capture the image and wait for the rest of the process to complete, then obtain the final result.
