Robot vision System that makes the robot have visual perception is one of the most important parts of the robot system. At present, it is widely used in the fields of electronics , automobile, machinery and other industrial sectors as well as medical and military fields. For the design of robot vision technology, there are several difficulties:

First, the stability of lighting

Industrial vision applications are generally divided into four broad categories: positioning, measurement, detection and identification, of which the measurement of the stability of the highest light requirements, as long as the light changes occur 10-20%, the measurement results may be deviations of 1-2 pixels, This is not a software problem. This is a change in lighting that causes the position of the upper edge of the image to change. Even the most powerful software can not solve the problem. Therefore, it is necessary to eliminate the interference of ambient light from the system design point of view and ensure that the active lighting source Luminous stability. Of course, through the hardware camera to enhance the resolution is to improve the accuracy of anti-environmental interference a way. For example, the corresponding camera before the space size is 1 pixel 10um, and by increasing the resolution into a pixel 5um, the accuracy can be considered about double the approximation, the natural environmental interference increased.

Second, the workpiece position inconsistency

Generally do the measurement of the project, either offline testing, or online testing, as long as it is fully automated testing equipment, the first step is to do the first thing to be able to find the test target. Each time the test object appears in the shooting field of view, to be able to accurately know where the test object, even if you use some mechanical fixtures, it can not be particularly high precision to ensure that the test object appears in the same place each time , Which need to use the positioning function, if the positioning is not accurate, it may not measure the position of the measurement tool is accurate, the measurement results sometimes have a larger deviation.

Third, calibration

Generally need to do the following calibration when high-precision measurement: First, the optical distortion calibration (if you are not using software lenses, generally must be calibrated); Second, the projection distortion calibration, that is, because you install the position error representative Image distortion correction, calibration of three-object image space, which is specifically calculated for each pixel corresponding object space size.

However, the current calibration algorithms are all based on planar calibration. If the physical to be measured is not planar, the calibration will need some special algorithms to deal with. Usually, the calibration algorithm can not be solved.

In addition, some calibration, because the calibration plate is not used, but also must design a special calibration method, calibration may not be able to solve all the existing calibration algorithm in the software.

Fourth, the object's speed of movement

If the object to be measured is not stationary, but rather in motion, then motion blur must take into account the accuracy of the image (blur pixel = object motion speed * camera exposure time), which is not what the software can handle.

Fifth, the measurement accuracy of the software

The accuracy of the software used in measurement applications can only be considered in 1 / 2-1 / 4 pixels, preferably 1/2, not 1 / 10-1 / 30 pixel accuracy for positioning applications because in measurement applications There are very few feature points the software can extract from the image.

Stone Railing

Stone Railing,Custom Stone Railing,Stone Railing Custom,Customization Stone Railing

Jinxin Stone , https://www.shandonghxs.com