Photogrammetry is gaining popularity in geotechnical engineering. Many researchers used photogrammetry to measure the soil volumes in laboratory tests. However, the potential of using low-cost security cameras in measuring the volume changes of unsaturated soils during triaxial testing has been largely unexplored. This paper presents a feasibility study on developing a multi-security camera photogrammetric method to reconstruct the 3-D models of the soil sample and track the continuous deformations of soil specimen during triaxial testing. Sixteen low-cost security cameras are installed to the steel ring around the triaxial test equipment. Coded targets are posted on the surface of the acrylic chamber and frame rods. The membrane is printed with coded targets. There are two stages for the proposed method. The first stage involves performing single camera-based photogrammetry before the triaxial test. A MATLAB program is developed to detect the coded targets automatically. Then, these coded target detection results are fed to a three-dimensional reconstruction program to determine the three-dimensional points in the air. The stations for each security camera can be back calculated. These camera stations will be used in the next stage for three-dimensional reconstruction. After that, multiple ray tracings are applied to determine the three-dimensional points on the soil specimen under water. The second stage for the proposed method involves multi-camera-based photogrammetry during the triaxial test. The security cameras record the whole testing process synchronously from multiple locations and orientations. Image processing is used to track the centroids of the coded targets. Ray tracings are applied again to determine the three-dimensional points on the soil specimens during the triaxial testing. Experimental results indicate that the proposed method can obtain accurate 3-D model of the soil specimen using relatively low-cost security cameras. In addition, the proposed method takes videos of the whole triaxial test, which can track the specimen deformations at every moment during the test.