Image Processing based Vehicle Identification and Speed Measurement

Vision‐based vehicle speed estimation: A survey

“…Second, those methods focused on tracking the centroid of the blob or the bounding box of the vehicle [25, 34, 35, 40, 48, 50, 52, 56, 65, 72, 73, 75, 83, 86, 95, 103, 106, 111, 118, 121, 122]. Finally, those approaches that focus on tracking the entire vehicle [15, 16, 18, 26–30, 32, 36, 42–44, 47, 54, 57, 63, 78, 81, 84, 89, 91, 98, 102, 104, 105, 108, 109, 111, 113, 114, 120, 125, 128, 129], or a specific part of it (such as the license plate [41, 49, 60, 66, 68, 70, 76, 77, 80, 82, 94, 100, 125]).…”

“…The first categorization is directly related to the attempt to devise an accurate system to be used for speed enforcement ( speed camera application domain) [13, 17, 18, 32, 33, 39, 41–43, 51, 53, 54, 61, 65–68, 70, 72, 74, 76–78, 80, 82, 84–86, 94, 100–102, 107, 108, 111, 112, 114, 119, 122–125] or a less accurate approach that can still be used for traffic speed estimation ( traffic camera application scenario) [12, 14–16, 19–31, 34–38, 40, 44–50, 52, 55–60, 62–64, 69, 71, 73, 75, 79, 81, 83, 87–93, 95–99, 103–106, 109–111, 113, 115–118, 120, 121, 126–129]. Although in some cases it is claimed that accurate speed detection is targeted, some aspects such as the low resolution of the camera, a limited experimental evaluation or the absence of an appropriate ground truth make us to categorize these manuscripts as traffic speed systems.…”

“…Other approaches are feature‐based, for example, detecting the license plate [31, 41, 49, 60, 66, 68, 70, 76, 77, 80, 82, 94, 122, 125] or other features of the vehicle [38, 39, 45, 49, 57–59, 69, 71, 79, 87, 91, 95, 96, 98, 99, 110–112, 116, 117, 123]. The use of learning‐based methods to recognize vehicles in images has been growing recently [19, 85, 88–91, 98, 100, 101, 104–106, 115, 120, 126, 127, 129].…”

“…Prior knowledge about the frame rate of the camera or accurate timestamps per each image are needed to compute the time between measurements. The use of consecutive [14, 18, 19, 22, 25, 33, 35–37, 41, 46, 48, 50, 55, 65, 67, 68, 71–76, 78, 82, 83, 85, 88, 89, 92, 96, 100, 103, 120, 129] or non‐consecutive [15, 24, 26, 28, 30, 31, 43, 44, 47, 49, 52, 59, 66, 70, 71, 86, 87, 93, 95, 99, 106–108, 111, 118, 124, 125, 128] images to estimate speed is a fundamental variable that has a considerable impact on accuracy. How to integrate all available measurements (instantaneous, mean, optimal etc.)…”

“…Surprisingly, and contrary to the evidence described in subsection 6.3, most of the works propose to use consecutive measurements (at frame and at frame ) to calculate the speed values for each vehicle [14, 18, 19, 22, 25, 33, 35–37, 41, 46, 48, 50, 55, 65, 67, 68, 71–76, 78, 82, 83, 85, 88, 89, 92, 96, 100, 103, 120, 129]. As illustrated in Figure 12, this technique tends to provide noisy values since the distance errors are of the same order of magnitude as the distance traveled by the vehicle.…”