Nowadays, the tensile test is undoubtedly the most essential material test for determining the mechanical properties of materials, both at an educational and commercial level. However, the universal testing machine is not completely accurate when acquiring the deformation data, presenting values greater than those existing in the interest of the specimen. For this reason, this project has the objective of developing a measurement system methodology for tensile tests based on image processing for metal and plastic tests, detailing each of its stages, from the capture and calibration of the images to the spatiotemporal synchronization of the recorded data with its tension/stress correlation or load level. Open-source programs such as Fiji and Python were used for image processing and data analysis, respectively, and a web camera for image capture. For the validation of the methodology, tensile tests of 4 metal specimens and 2 plastic specimens were carried out at ESPOL laboratories at different capture rates, where the parameters of the tensile tests were obtained from the ASTM E8 and ISO 6259 standards. For these tests, stress-strain curves were constructed that faithfully represent the typical tensile behaviour of each material, reaching higher precision levels compared to the results provided by the testing machine. It is concluded then that a systematic methodology of measurement by images for tensile tests is feasible and can also capture the characteristic behaviour of tensile materials. It is also a low-cost alternative, whose total value is 3% of the cost of its equivalent accessory on the market.