What are DVT Cameras used for in Manufacturing?

DVT cameras also referred to as DVT vision sensors are smart image cameras developed by DVT Corporation–a global manufacturer of general-purpose machine vision systems, offering the widest range of vision sensor products for use in the manufacturing industry. These smart image cameras from DVT are simply nodes on a factory network that collect and process visual data to motion controllers and robotics, perform part inspection, report part inspection results for quality analysis and defect detection, and which carry out inventory control and tracking of parts on an assembly line. They operate with Intellect Software which facilitates their setup, configuration, integration into a machine vision system, and maintenance. They are also backward (downward) compatible with API (Application Programming Interface) FrameWork.

DVT cameras have integral Ethernet for connecting to Intellect Software and factory floor networks, allowing the setup of network communications with other vision systems (i.e. computer vision systems), Programmable Logic Controllers (PLCs), databases, personal computers (PCs), and Supervisory Control and Data Acquisition (SCADA) systems.

The Intellect Software version available as an Intellect CD for operating DVT cameras should be installed on a PC with the following specifications:   

• A High-Performance Processor with faster processing speed for faster updating of captured images into the Intellect user interface. 
• At least 128 MB RAM: In DVT cameras, all processing is done in the RAM memory. Also, almost all real-time data including new product features reside in RAM. 
• 16 MB Flash Memory: DVT camera data that is required to be persistent–data that should not be lost upon power up and power down of the camera–is stored in the flash memory. 
• Space Availability: Intellect Software requires minimum hard disk space of 100 MB. 
• Ethernet Card: To properly install the Intellect Software, your PC should have a 10/100 Mbps Ethernet card with an integrated RJ-45 connector jack. But built-in Ethernet is highly recommended. 

Using standard Ethernet network communications, DVT cameras support several industrial network protocols including EtherNet/IP, Modbus/TCP, and TCP/IP. Also, these cameras can support DeviceNet and PROFIBUS protocols through an optional Smart Link module. The RJ-45 connection on the DVT camera is designed to establish communication between the DVT vision sensor and the PC running the Intellect Software.

History of DVT Cameras

DVT Corporation invented the first smart image camera in 1990, and in 1998 it introduced Ethernet capabilities into its smart cameras. Later on, in 2001 it produced the first smart image camera with mega-pixel high resolution, and in 2002 it developed a DVT Spectral vision sensor that became the first smart image camera to use a prismatic optics system for high-resolution and extremely precise color measurements.

Recently, DVT Corporation launched a fully functioning, extreme-speed 55X smart camera series with a 4000 MIPS (Million Instructions Per Second) Digital Signal Processor (DSP). Despite being the newest entry in the vision sensor market, the 55X Series is readily affordable at a price of less than £2,000.

Current DVT cameras can be considered as self-contained machine inspection systems that provide a visual data source with high speeds, low network connectivity costs, seamless data exchange within a smart manufacturing enterprise and enhanced vision tools that support industrial automation locally and globally.

DVT Cameras in Machine Vision

Machine vision can be described as the ability of a computer to visualize objects or scenes. This technology employs one or more smart image cameras or vision sensors like DVT cameras, Analog-to-Digital (A/D) converters, and Digital Signal Processors (DSP) to acquire data that are then sent to a robotic controller or computer (PC) to control a manufacturing process or equipment.

Machine Vision is occasionally confused with Computer Vision. But while machine vision focuses on the most crucial parts of the captured image relative to its application to determine appropriate action like automating tasks, computer vision often zeroes in on understanding the image fully after capturing, processing, and analyzing it. For this reason, computer vision technology is oftentimes integrated with Machine Learning (ML), Deep Learning (DL), and Artificial Intelligence (AI) algorithms to accelerate image processing and interpretation in order to extract meaningful information. In terms of complexity, machine vision is similar to voice recognition technology whereas computer vision tries to replicate the human vision system.

In essence, machine vision can be thought of as a subclass of computer vision. This is because a machine vision system applies vision sensors to view an image and computer vision algorithms to process and interpret the captured image, before using the processed data as a digital input to instruct other automation components in the system to act upon that data. But computer vision does not need to be part of a larger machine system, as it can be used alone to create digital systems that replicate complex parts of the human visual system.

How Does a Machine Vision System Work?

Machine vision systems make use of multiple DVT cameras, processing hardware, and computer vision software algorithms to automate complex or routine visual inspection tasks in an industrial or manufacturing environment. The DVT cameras capture images from the surrounding environment, while the combination of the processing hardware and software algorithms process the acquired visual information, preparing it for use in various manufacturing applications.

The DVT cameras in a machine vision system use specialized optics to acquire the images. This approach lets specific characteristics of the acquired image (critical parts of the image) be measured, processed, and analyzed, which is the basis of machine vision technology. For example, if a machine vision system with DVT cameras is integrated with a manufacturing system, it can be used to analyze a particular attribute of a product being processed on an assembly line. It can then determine if the product meets the pre-defined quality criteria, and if not, it will inform other automation components in the manufacturing system to dispose of the faulty product.

Typical machine vision systems applied in manufacturing settings require the following components:   

• Lighting: Used to illuminate the manufacturing scene or the part being manufactured to make its features visible. 
• DVT Camera(s): The DVT camera provides a lens and a sensor for capturing and digitizing the image. The lens captures the image of the required object and delivers it to the DVT sensor in form of light. The sensor converts the light into digital electric signals using an embedded charge-coupled device or complementary metal-oxide semiconductor technology. Thus, the image captured by the lens is converted into a digital format as pixels. 
• Processor: It runs computer vision software algorithms to process the digital image from the DVT camera and extract the required information. The processor can consist of a PLC, robot controller, or an industrial PC with Intellect Software containing machine vision algorithms, Artificial Intelligence (AI), and middleware software algorithms for data analytics and processing. 
• Output: The machine vision system output makes use of the data from the processor to control robots and communicate information to other manufacturing system machines or automation software. Output systems enable the DVT cameras and processing system to communicate with other elements of the manufacturing system (i.e. conveyor belts, robotics, etc.), usually by physical discrete I/O connections or through serial communications. 

Role of DVT Cameras in Manufacturing 

In a manufacturing setup, DVT cameras are your hardware tools for gathering data, assuring the quality of products, and providing information on various manufacturing operations to all levels of your enterprise. This information can be categorized as process-level control data like SPC (Statistical Process Control) data or device-level operating data such as inspection values that are transmitted to a PLC.

The shared information can also be generalized as a FAIL or PASS signal or as specific as a feedback signal for the actual motor position being sent to a motion controller. The Intellect Software operating DVT cameras in manufacturing systems is used to divide and conquer the aforementioned tasks and is designed for ease of use by anyone.

Outlined below are the key functions of DVT cameras in manufacturing. Note, that DVT cameras in manufacturing applications are often integrated into machine vision systems rather than being used as stand-alone devices.

A) Defect Detection 

Detection of flaws in parts being manufactured happens to be one of the most essential quality control tasks in manufacturing facilities and is also the most utilized function of DVT cameras. In flaw detection, the DVT cameras search for specific defects such as scratches, cracks, discoloration, gaps, blemishes, or contaminants on the surface of the part(s) being manufactured. They detect flaws that can affect the reliability and functionality of the end product.

Such flaws appear randomly, so the software algorithm running the DVT camera looks for changes in texture or color, pattern changes, connected structures, or discontinuities. It then monitors the presence of any of those defects. And if a defect is detected, it’s classified in terms of type, size, color, and texture. Next, the machine vision system containing the DVT camera instructs the manufacturing system to dispose of the defective parts that don’t meet the specified quality criteria.

DVT cameras in manufacturing systems can readily and effectively detect minute and microscopic defects that are invisible to the human visual system. They can also operate tirelessly for prolonged periods of time, unlike human inspectors who get tired easily.

Flaw detection using DVT cameras is widely used for quality control in the manufacturing of electrical appliances, semiconductors, electronic components, materials produced in continuous rolls (such as metals, paper, and plastics) as well as food products and their packaging. It’s also useful in online inspections of tooling conditions and manufacturing processes, where a faulty process is halted immediately and the operating parameters are corrected while the defective parts are separated from the rest.

B) Verifying Assembly Line Integrity

A typical application of DVT cameras is on an assembly line, where after a given manufacturing operation is performed on a part, the DVT camera is triggered to capture and process the part’s image. Since DVT cameras are smart image cameras, they can be programmed to check the position of the part, its shape or size, and whether the part is present on the assembly line or not.

Once the part has been inspected as per the user-defined program, the DVT camera generates a signal to instruct the assembly line on what to do with that part. The part can be rejected onto an offshoot conveyor belt or into a container or passed on to undergo more assembly operations with the machine vision system tracking its inspection results (as inspected by the DVT cameras) throughout the manufacturing system. DVT cameras also precisely guide the handling equipment during the assembly of the end product.

Essentially, DVT cameras on assembly lines provide a lot more information about a part than simple proximity sensors that only detect the presence or absence of an object. For example, they allow manufacturers to inspect bottles on an assembly line in a full 360° view, ensuring that the finished products are placed in the correct packaging. The DVT cameras can also inspect other critical attributes of packaged end products such as labels, cap seal/closure, print quality, position on the assembly line, etc.

C) Counting/Sorting

This is a type of presence inspection in which DVT cameras confirm the presence or absence of parts and their quantity. It’s one of the most widely performed tasks in most manufacturing industries and a basic operation carried out by machine vision systems with DVT cameras. Practical applications of this function include counting countable finished products such as screws, bottles, automobile parts, jewelry, etc., and checking the presence of labels on electronic components on Printed Circuit Boards (PCBs), washers/screws in fastened parts, food packaging and adhesive application.

D) Practical Positioning

This is the process of comparing the actual location and orientation of a certain part to a specified spatial tolerance. The DVT cameras communicate the location and orientation of a given part in 3D or 2D space to a robot controller or another machine element in the manufacturing system, for the latter to place the target object in its proper location and align it in the correct orientation. The use of DVT cameras in machine vision provides more accurate and high-speed positioning systems compared to manual inspection, positioning, and alignment.

Practical positioning applications of DVT cameras in manufacturing include checking of barcode and label alignment, robotic arm pick-up and placement of parts on and off conveyor belts, arrangement of finished parts packed in a pallet, checking of Integrated Circuit (IC) placement in a PCB, and positioning of glass substrates in a matrix mold.

E) Part Measurement

DVT cameras can check the geometric tolerances and dimensional accuracy of parts by calculating the distances between two or more specified points and the location of the targeted features on the captured image, so as to determine if the measured value is within the specified range. To accomplish this, DVT cameras are used in a machine vision system that optimizes its lighting and optical components in order to obtain precise, repeatable, and highly accurate measurements.

DVT cameras can measure dimensional features as small as 25.4µ (microns). This function is normally combined with defect detection to measure the deviations of the detected irregularities on a part(s). It is also used to calculate the volume of the parts being manufactured.

F) Part Identification

DVT cameras used in machine vision for manufacturing systems can scan and decipher 2-D or standard barcodes, direct part marks, or even read printed characters on labels, packages, and parts. On reading such barcodes, markings, labels, etc. the DVT cameras can assist in part identification as the former contains the product name, date code, expiration date, lot number, and manufacturer. Part identification in a manufacturing setup is useful in improving inventory control, traceability of parts, and the verification system of finished products.

Note: By tracking parts on an assembly line, eliminating defective parts, reducing production runtimes, improving inventory control, and facilitating compliance of finished products with regulations (quality control) using DVT cameras in machine vision systems, manufacturers can increase production yields, save on production costs and increase profitability.

This entry was posted on August 31st, 2022 and is filed under Automation, Education. Both comments and pings are currently closed.