Want to play with the machine vision system _ you have to understand these top ten problems

To provide guidance to users on the key, fundamental characteristics that should be considered when choosing a machine vision system. Here are ten areas to prioritize when choosing a machine vision system:

1 positioner

The precise positioning of an object or feature is an important function of a detection system or a visually guided motion system. Conventional object positioning uses gray value correction to identify objects. Although this technique has been widely used, it lacks stability in the case of poor image quality. Image quality degradation may be due to factors such as clutter, brightness, and occlusion. In contrast, geometric object localization is the most up-to-date method of using objects' outlines to identify objects and their features. Both Adept's HexSight software and the AdeptVision sAVI system use very advanced contour detection techniques to identify objects. A reliable positioner saves cost by eliminating the need for fixtures to position parts.

2 light and lighting

It is well known that proper lighting and lighting are critical to the success of visual applications. Sometimes, although the most appropriate lighting and lighting are selected, if the object or feature being monitored is slightly changed, the lighting and lighting brightness are required to be changed accordingly. For example, the polished surface of the chip differs in the quality of the laser etched mark in OCR applications. A well-stabilized positioning tool such as Adept's geometric object locator can easily handle situations where image quality deteriorates due to changes in light accumulation and brightness. Using the right software allows you to eliminate the necessary dimming operations, regardless of how their image quality changes.

3 complete toolset for multiple tools

Machine vision software is sold in two typical forms—one for a complete set of visual tools and one for tools for specific tasks, such as BGA detection. The end user's application will decide whether to use a complete set of visual tools or a number of specific tools. A visual tool is a general application or algorithm that performs a predetermined task on an image or a portion of an image. For example, a spot detection tool can find a set of dark or bright pixels and measure the various sizes of the spot. The focus of choosing a vision system is to have a complete set of visual tools. Although you don't need all the tools at the beginning of the project, your requirements will change in a few years, and additional tools may be needed for new applications. Adept's HexSight software is a comprehensive machine vision library based on today's popular software platforms such as Visual Basic, Visual C++ and AcTIveX controls. The optional HexSight symboling tool is available with OCR, digital matrix and barcode decoding tools. The AdeptVision sAVI system is delivered with a simple graphical interface for visual and motion application programming.

4 programming and operation convenient and simple, intuitive graphical interface is the key to easy to use and set

The main difference between today's machine vision products is their graphical interface. The interface should be evaluated from both "setup" and "operation". For an engineer, it should be very complicated and should be very simple for an operator. For example, the calibration of a vision system should be as simple as a few clicks of the mouse, rather than as an exercise in rocket science.

5 sub-pixel precision

The resolution of the vision system is the smallest feature that the system can distinguish. For example, a visual range (FOV) of '1' would give a resolution of 1/640 or '0.00156' using a computer image of 640 x 480 pixels. In fact, machine vision algorithms have sub-pixel capabilities. That is, these algorithms can measure or derive smaller units than one pixel. Adept's geometric object locator tool can be accurate to 1/40 of a pixel. All other vision system tools can only be accurate to 1/10 to 1/20 of a pixel. The subpixel accuracy of a vision tool depends on the quality of the image and the power of the algorithm.

6 future upgrades

Machine vision systems can be used in a variety of applications, ranging from trendy cameras to surveillance systems. Consideration should be given to future upgrades to the system when selecting a system. Universal vision software from several vendors provides the end user with the right lighting, optics and vision tools. Dedicated software packages such as BGA detection, OCR, etc. can also be sold as pre-configured software. A vision software system based on general purpose is better upgraded. End users should consider future system requirements based on additional cameras, changes in lighting, changes in visual tools, and more. For example, if a system that requires multiple cameras is to compare the price and flexibility of a frame-based acquirer-based system with a trendy camera-based system. Adept's HexSight software is compatible with a wide range of third-party frame grabbers and cameras, including FireWire cameras.

7 image preprocessing

It is very important to detect feature points and defects, regardless of the brightness and surface or material of the object. The image preprocessing algorithm can magnify the feature points of the image so that the visual tools can detect them better. Similarly, feature points can be scaled down so that visual tools can ignore them. For example, HexSight's morphological operations can be used to remove or fill small holes in an object, to separate objects at a sparse point or to connect adjacent objects. Similarly, the filtering operation can be used for convolution of the input image. The availability of such a wide range of pre-processing algorithms is key to the success of complex vision applications.

8 visually guided movement

If your application requires a vision system, you must know how the vision system is integrated with the motion system. For calibration and operation, the unintegrated motion system and vision system are preliminary systems, and the robot or mechanism and vision system are separately calibrated. In operation, a separate vision system calculates the offset of the part position based on the known position in the visual coordinate system, and then issues instructions to the robot's arm to pick up the part at an offset from the initial programmed pick position. . Instead, an integrated system such as the Adeptvision sAVI controller calibrates the vision system and machine in a coordinate system. The part positioning can then be defined in the same six-degree-of-freedom coordinate space as programmed in the machine.

9 system integration

If you are not very proficient in machine vision technology, then you need a system integrator for your project. The ideal visual product is widely accepted by system integrators. Adept has a program called Adept-Priority Visual Integrator (PVI) that trains system integrators on Adept products. These system integrators have application engineering resources to handle the toughest machine vision projects. In addition, these PVIs are ideal for handling Adept's robotics and sports products. This is an ideal resource for projects that require motion and vision.

10 factory floor connection

At present, there are various methods for communicating with the visual system. The common interfaces are serial port (RS-232), RS-485, parallel port, Ethernet, Devicenet, digital I/O, and the like. Newer interfaces such as IEEE-1394 and USB are also widely used. When evaluating a vision system, consider the connectivity of the plant floor. Typically, a machine vision system is a data acquisition system that interfaces with information systems of other plant floor equipment and plants. Some vendors' software can remotely operate the vision system over a local area network or the Internet. In certain industries, such as pharmaceuticals, machine vision systems are required to be independent for a particular application, ensuring that settings are not disturbed by remote operations. When choosing a vision system, the communication interface of the vision system is an important consideration and should not be ignored. The AdeptVision sAVI controller can interface with Rs-232, Rs-485, Devicenet, Ethernet and Firewire interfaces.