Inspired by the tactile sensing abilities of humans, tactile tactical sensors are used in many areas from robotics to heavy industry, such as surface roughness determination, softness determination, and quality control, etc. in industry. In the first stage of this study, for the production of a tactile tactical sensor from the silicon family, the most suitable tactile tactical sensor mixing ratios were determined according to the purpose of use by making softness, drying time, strength, flexibility, and light transmittance depending on the mixing ratios. In the second stage, the surface roughness images of the samples with known dimensions were taken by using the tactical tactile sensor produced with the determined mixing ratio for the determination of sub-millimeter surface roughness, in the microscopic imaging setup. In the third stage, image processing software was made with the original algorithm created using the MATLAB program, and depth detection calibration from the image was performed. As a result of the experiments, it has been determined that tactical tactile sensors and image processing software provide realistic depth information with 2-micron sensitivity and can be used for surface roughness. Inspired by the tactile sensing abilities of humans, the tactile tactical sensors are used in many areas from robotics to heavy industry, such as surface roughness, softness determination, quality control, etc. in industry. It is needed for purposes. In this article, it is aimed to develop a tactile tactical sensor and imaging system for artificial intelligence-based detection of submillimeter surface roughness. The developed system consists of three stages; In the first stage, the production of silicone rubber, which forms the tactile tactical sensor, is emphasized. Silicone rubber is obtained by reacting three different chemical components, namely A, B and C. The varying ratios of these chemical components give us different characteristics such as softness, drying time, strength, flexibility and light transmittance. In order to achieve the best tactile feature for the sensor, the optimum mixing ratio of the silicone rubber components that make up the tactile tactical sensor was determined by the Taguchi method. In the second stage of the system, the tactical tactile sensor, produced with the ideal mixing ratio determined for the determination of sub-millimeter surface roughness, was used in the microscopic imaging mechanism, and surface roughness images of the samples with known dimensions were taken. In the third stage of the system, image processing software was made with the original algorithm created using the MATLAB program, and depth detection calibration from the image was performed. As a result of the experiments, it has been determined that tactical tactile sensors and image processing software provide realistic depth information with 2-micron sensitivity and can be used with high reliability in surface roughness. This study allowed the surface depths of images taken with the sensor to be calculated on a pixel basis. With this tactile sensor, which does not take up space and has high mobility, depth analysis of hard-to-reach surfaces can be done easily and precisely. In addition, the tactile sensor can be considered a very useful apparatus in robotics technology for providing the robot with a sense of touch.