Tactile pattern super resolution reconstruction method and acquisition system

A tactile pattern SR acquisition system, comprising:

• a robotic arm, a tactile sensor, and an upper Personal Computer (PC), wherein
• the tactile sensor is mounted at an end of the robotic arm;
• both a control end and a signal output end of the robotic arm are connected to the upper PC;
• the upper PC is used for receiving the coordinates of the center point of the tactile sensor and controlling the robotic arm to move according to a preset acquisition path according to the coordinates of the center point of the tactile sensor;
• a signal output end of the tactile sensor is connected to the upper PC;
• wherein the upper PC comprises a memory which stores instructions and a processor, wherein
• when the instructions are executed, the upper PC is configured to perform following steps:
• obtaining tactile data and coordinates of the center point of a tactile sensor at each sampling of each contact surface;
• registering all sampled tactile data of each contact surface according to the coordinates of the center point of the tactile sensor at each sampling, so as to obtain an HR tactile pattern sample corresponding to the contact surface;
• selecting the tactile data sampled each time in a central area of the contact surface and the HR tactile pattern sample corresponding to the contact surface to form a data pair, and constructing a tactile SR data set;
• training a deep learning neural network model by using a supervised machine learning method according to the tactile SR data set, so as to obtain a tactile SR model;
• and
• inputting the tactile data, acquired by the tactile sensor, of a contact surface to be measured into a tactile SR model, and outputting an HR tactile pattern of the contact surface to be measured,
• wherein the registering all sampled tactile data of each contact surface according to the coordinates of the center point of the tactile sensor at each sampling, so as to obtain an HR tactile pattern sample corresponding to the contact surface specifically comprises:
• setting acquisition times I along an X axis and acquisition times J along a Y axis of a robotic arm;
• averaging the plurality of sets of tactile data acquired at each sampling after high pass filtering, and taking an average value as the tactile data of each sampling;
• registering by using a formula THR [I×m+i, J×n+j]=Ti,j [m, n] according to the tactile data of each sampling of each contact surface and the coordinates of the center point of the tactile sensor, so as to obtain an initial HR tactile pattern sample, wherein [m, n] represents the [m, n] th taxel of the tactile sensor, THR [I×m+i, J×n+j] represents the x, y, and z-axis data corresponding to the [I×m+i, J×n+] th pixel point of the HR tactile pattern, represents the x, y, and z-axis data corresponding to the [m, n] th taxel when the tactile sensor samples for the ith time along the X axis and samples for the jth time along the Y axis;
• and
• performing smoothing processing on the initial HR tactile pattern sample to obtain an HR tactile pattern sample of the contact surface.