Abstract Automatic distance and dimensions estimation of pedestrians are sometimes imperative in real-time scenes. Such estimations are needful when contact based measurements are unrealistic. It is desirable to have a non-contact measurement framework. This work exhibits a method that obliges simple mathematical estimations to automatically discover the distance and dimensions (height and width) of a moving pedestrian lying at distant locations from the camera. The proposed system confines to immovable monocular camera environments. The foremost step before measurements is a single-shot environment learning. An L-shape marker is used and its cornered points are detected and stored by placing it first at a minimum distance and then at a relatively far distance from the camera. At the two placements, the cornered points of the marker are deemed to be joined by four straight lines. With the help of line equations, per-pixel-length of object's location is calculated. The mean filter is then applied for background subtraction to extract foreground objects. Pedestrians are then classified by passing foreground objects to a convolutional neural network based classifier. Afterward, distance is calculated with reference to the smallest known distance between a pedestrian and the camera. Thereafter, the approach estimates height and width of the pedestrian. Outcomes are compared to the found results of existing methods as well as with the real measurements. The results show vigor of the proposed framework with worthy lapse rates.