Can computer vision and remote control be added to an analogue measuring instrument? In his master’s project, Ivan Zahorulko developed an automated data acquisition system for the GNU-KS gravimeter. It combines electronics, programming, image processing, and precision mechanics to automate measurements and reduce the dependence of the results on the operator’s actions.
What problem does the project address?
A gravimeter measures gravitational acceleration. These measurements are used in geophysical and geodetic research. When working with an analogue gravimeter, the operator observes a light marker through an eyepiece, manually turns a micrometer screw, and reads the indications. The result depends on the operator’s attentiveness and the precision of their actions. The sensitive instrument may also be affected by heat and vibrations associated with the operator’s presence.
The project’s objective is to automate these operations and enable the instrument to be operated remotely while preserving its sensitive mechanical system.
How the system works
Two digital cameras monitor the gravimeter’s light marker and readout mechanism. A Raspberry Pi 4 single-board computer processes the images and coordinates the system components. The computer vision algorithm isolates the required image region, filters out noise, and determines the position of the light marker. Based on these data, the system controls a stepper motor that turns the micrometer screw and returns the sensing element to a specified position. The operator can remotely monitor the instrument and control the system over a network.
What has been developed?
The project involved designing and assembling a system comprising:
- a Raspberry Pi 4 Model B for data processing and control;
- two USB cameras for monitoring the instrument readings;
- a stepper motor and driver for controlling the micrometer mechanism;
- custom-designed mounts and an enclosure manufactured using 3D printing;
- an LED illumination system;
- software for image processing, control, and remote interaction.
The software was developed using Python and the OpenCV library, with FastAPI used for the web interface. The enclosure was modelled in Blender.
How the system was tested
The system was tested under laboratory conditions. For comparison, two sets of 30 measurements were performed: one with manual readings and one in automatic mode.
According to the results presented in the thesis, the standard deviation was:
- 0.029 mGal in manual mode;
- 0.006 mGal in automatic mode.
Thus, the spread of the results decreased by a factor of approximately 4.8. This indicates improved measurement repeatability in the laboratory experiment performed.
The student’s contribution
Ivan Zahorulko analysed the gravimeter’s operating principle, defined the technical requirements, and selected the components. He developed the system’s hardware and software, implemented computer vision and motor control algorithms, and carried out experimental studies and processed the measurement results. The project brought together practical skills in working with electronic components, programming, 3D modelling, automatic control, and data analysis.
Practical significance and further development
The development demonstrates the potential to modernise analogue scientific equipment using digital cameras, an embedded computer, and software. This approach may be useful for research laboratories and educational measurement systems.
Further development will focus on increasing autonomy, improving recognition algorithms, adding georeferencing of the results, and adapting the system for field measurements.
Author: Ivan Zahorulko, second-year master’s student, group DS-41mn, educational and research programme “Electronics”.
Academic supervisor: Ievgen Kovalenko, Candidate of Technical Sciences, Associate Professor.
Year: 2026.
Address: 16, Politekhnichna St., 03056, Kyiv, Ukraine
Phone: (044) 204-90-70, E-mail: kafedra-eds@lll.kpi.ua
