The specific process is as follows: the image captured by the camera-> enters the DSP/PC platform-> gray image-> gray display; Binary image->; Determine the boundary (expansion/contraction algorithm can be used)->; Calculate the enclosed area-> Identify part types according to features-> Calculate the number of parts.
There are two kinds of papers published in this direction:
1. algorithm simulation class, according to the collected test images and relying on Matlab simulation tools, design an algorithm with excellent performance for one or several parts (such as boundary determination part and part feature recognition part, etc.). ).
2. In terms of hardware, based on DSP or PC platform, the images collected by the camera are analyzed and processed in real time, and a practical recognition system is designed. This is very difficult and challenging.
Most of the mechanical majors I know and those of us who are job-seekers are engaged in hardware research, because the tutor's involvement in the field of communication/automation is the need of the project, but interdisciplinary research cannot be completed by one student. I don't know what your tutor wants from you, but you can describe it in detail.
I hope I can help you. You can ask me again if you have any questions.
-Technical Department of China Internet of Things School-enterprise Alliance