Surface recognition improvement in 3D medical laser scanner using Levenberg–Marquardt method

Improve 3D laser scanner measurements accuracy using a FFBP neural network with Widrow-Hoff weight/bias learning function

“…40 to improve the convergence speed. Addi− tionally, the presented laser scanner is able to change its scan− ning frequency in a certain range, and as shown in our previous experimentations [34,35,41], at various scanning frequencies some characteristic parameters of the scanner functioning are changing. However, it does not affect measuring uncertainty in experimentation with static objects, and in general behaviour it is congruent with the data of our Table 3.…”

“…The Widrow−Hoff learning function was used to train a neural network and adjust the scanner measurements, working with the Levenberg−Mar− quardt training algorithm [34,35] this function shows a reli− ability of 99.97% in the measurement adjustment. The accu− racy of the adjustments, even out of the normalized errors (as in Fig.…”

An approach for dynamic triangulation using servomotors

“…At the institute of engineering UABC the technique of triangulation was extended to the dynamic triangulation [4] - [9]. The principle of dynamic triangulation varies the emitter (PL) and receiver (SA) angle, to increase the field of vision: It was developed the prototype of laser scanning No.2, which by stepping motors varies the angles of the emitter and by a DC motor the angle of the receiver.…”

“…Triangulation is the classic method for land surveying in geodesy and consists in division and measuring an area by triangles. These triangles form a mesh, where angles and distances are calculated using trigonometric functions [4]. To measure distances using laser triangulation, be considered a generally triangular in Figure 1, where the distance is searched.…”

Transient signal analysis based on Levenberg–Marquardt method for fault feature extraction of rotating machines

“…parameter identification for each transient. In each iteration, transients are modeled by a parameterized double-side asymmetric one based on Morlet wavelet, and then the Levenberg-Marquardt (LM) method [47][48][49], as a nonlinear least square method, is introduced for parameter identification. With the implementation of the iterative extraction, transients in vibration signals can be analyzed and extracted one by one.…”