Special delivery

Share this on social media:

Topic tags: 

With Christmas fast approaching, staff at distribution hubs of retailers as well as the postal service will be working exceptionally hard. Tom Eddershaw investigates how machine vision can help automate processing packages at distribution centres to get presents delivered in time for Christmas

Imaging technology is set to transform transport logistics, in a shift that one observer believes is as significant as the move from analogue to digital technology. As consumer habits in the developed economies change, and shoppers increasingly buy from online retailers such as Amazon, any technique that can improve the efficiency with which their parcels are despatched will save costs and bring economic benefits.

The latest imaging technology will allow logistics companies to benefit from improved automated systems to diagnose problems with markings, to check the integrity of a product before it leaves the warehouse, and to ensure that the lorry is leaving with every inch of available space being used. ‘It’s an exciting time; it’s a whole new technology and a whole new way of tackling a problem,’ said Leigh Jordan, the UK ID product manager for Cognex. Replacing laser line scanners with imaging technology ‘parallels the analogue-to-digital evolution that we have seen everywhere else’. He explained that the image-based systems prove their worth when laser scanners, the ‘analogue’ technology, struggle with issues such as omni-directional scanning and damaged codes.

One of the primary uses for an image scanner is identifying products by a barcode on the conveyor belt. As well as having higher read-rates than the traditional laser scanners, image scanners are also capable of handling codes that are damaged, faded, of poor contrast, broken, warped or at extreme perspectives. Moreover, imaging systems, unlike lasers, can count pieces, measure the volume of an object, and discern its position on the conveyor belt in a single pass.

But laser sensors are cheap, and work to high degrees of accuracy, normally around 95 per cent. Image-based scanners run at 99.98 per cent, and during head-to-head testing have been reported to achieve read rates of even closer to 100 per cent – yet they require significant investment in re-engineering the systems in warehouses, something that costs both time and money. So does a 4.98 percentage point increase in successful reading of a barcode really make enough of a difference to warrant the reform?

According to Jordan, the simple answer is: ‘yes’. With the number of packages travelling through a distribution centre, that increase of 4.98 percentage points can make a difference to the number of read-errors made per day. This translates directly into economic and business efficiency; the fewer mistakes made, the more products that can get out of the door and on to the customer.

The way a laser system works is by shining a bar of light over a barcode and collecting the light reflected by the white parts of the code. From this data, the system can calculate the light absorbed by the code and uses this to decide what the label relates to. The scanner uses mirrors to increase the amount of bars of light that may come into contact with the barcode – but, if there is damage to the code or it’s too distorted, the laser system will not recognise the pattern. This doesn’t occur with image scanners, which capture a still image of the area of interest and then software algorithms identify the pattern from a small slither of the code. The systems are also capable of decoding 1D and 2D barcodes, whereas most laser scanners can only recognise the simpler 1D type.

Imaging systems, on the other hand, normally have a light to illuminate the product, cameras to take the image and a means to transmit the image to the central system for processing – normally an Ethernet or USB cable. The lens used can either be a standard or a liquid lens, so the camera will either be fixed and auto-focus, or have to move in order to create a focused image.

A series of cameras, typically six, is set up at a tunnel that the package or product passes through, allowing for complete coverage of all sides (including the surface in contact with the belt by means of a small break in the conveyor). This means that it doesn’t matter which way the package is placed on the line, reducing the required manpower. The cameras can either be fully independent or work in unison by means of an Ethernet cable and a central computer. They can be programmed to capture an image sequentially with gaps of the order of a millisecond to avoid one flash impacting another camera’s view. If they are independent, multiple cameras can be used to process images with no added strain to the system as a whole and freeing up computer space.

However, the impact of imaging systems in the warehouse and transport logistics is not limited to more efficient ways of reading barcodes. Uwe Pulsfort of Teledyne Dalsa explained that they can also be used to complement laser technology by measuring the volume of a package, for example. He pointed out that if the amount of space available in the lorry is known, the systems can calculate the best way to fit the most packages into the truck, so that no space is wasted. This means fewer trips need to be made which can save money for a company every day.

The cameras can run at speeds of up to 120 frames per second, which allows multiple images to be taken as the part goes past. As the engineer can see where the sensor is looking, it is very quick and simple to install and calibrate the camera. The whole system can also give immediate information regarding a product’s location or condition, which means the products that are on the conveyor belt can be observed constantly if the operator requires it. The images that are taken can be stored, meaning that the condition of the product can be monitored as it leaves the warehouse, showing clearly whether anything was wrong before distribution, and acting as possible evidence if there is any dispute when it reaches the customer.

According to Pulsfort, the varying size, position and orientation of products or packages present challenges to the efficient operation of the image sensors. Different packages may be of different sizes or they may be placed in different positions on the conveyor – all of which changes the depth of field and means that the scanner has to change focus very quickly if it is to accurately image the area that is important. Some of the Cognex scanners support liquid lens technology which can auto focus very quickly. Liquid lenses use the difference in the refractive index of water and oil, and the fact that the liquids don’t mix, to adjust the focal length of the lens. However, many scanners have fixed lenses so, to focus the image, the scanner itself has to move.

A further area of application for imaging scanners to improve speed and efficiency in the warehouse is address recognition, either printed or handwritten. Optical character recognition (OCR) speeds up sorting for mail distribution and automates the system so it can run for longer, at a lower cost to the company. On the rare occasion of a misread in the German postal service, according to Pulsfort, an image is sent via the internet to a support team that analyses the image and sends the correction back, allowing the letter to return to circulation within seconds. This also means less staff are required to run the distribution centre, reducing cost, time and any risk to employees who work with fast moving machinery.

 

Image-based scanners are more accurate than barcode scanners, which saves time and money when sorting volumes of packages. Credit: Cognex 

In the future these scanning systems are looking to operate even faster, improve the read rates further, address the focusing issues and become cheaper. Their ability to serve multiple system needs simultaneously, though, makes them very valuable to a logistics company.   

Robotics and Industry 4.0

As warehouses lead towards Industry 4.0, the German-coined term referring to the fourth industrial revolution to make industry fully automatic, the use of robotics, and therefore also vision, is becoming more and more important in logistics to automate mundane and potentially dangerous tasks. Fetching products from storage for distribution (long hauling) or loading packages onto pallets, for instance, could both benefit from vision-guided robots.

A company called Seegrid, from Pittsburg, USA, makes Automated Guided Vehicles (AGVs) that can carry pallets around a warehouse, forklift trucks that can load pallets onto shelving systems or towing tractors to pull loads at speeds of up to 2.4mph for use in very large warehouses. The company is using imaging technology to help its products navigate their way around the warehouse.

Tom Bonkenburg, director of European operations for St Onge Company, a consultancy for supply chain and engineering within warehouses, explained how the process works. With the Seegrid range an operator walks the AGV through the warehouse in order to build up a visual blueprint of pre-programmed paths through the warehouse. An employee of the warehouse will then give the AGV a pallet or load to take to another part of the warehouse and chooses one of the pre-programmed paths for the robot to automatically take to the required area.  

The Amazon-owned Kiva systems of North Reading, USA, manufactures a smaller AGV, used mostly in the e-commerce industry, which uses a collection of cameras that read barcodes that are stuck to the floor as a fiducial marker, carrying information of the location. The robot drives from sticker to sticker to navigate to its destination. The robots receive further information from a central server to assist in the navigation, which helps the AGVs avoid collisions. When an employee requests a product to be packed and distributed, a Kiva robot collects the shelving unit that the object is stored on and brings it to the packaging area.

‘The benefit of this [use of imaging] is that you don’t need to start putting wires or magnets into the floor, you don’t need to start putting up reflectors that lasers can reflect off to triangulate position, which are the traditional ways AGVs navigate around the warehouse,’ Bonkenburg said. ‘Using imaging to navigate is a big advancement in the application of robotics in the warehouse environment.’

Companies are making these devices more independent, faster operating and cheaper as time passes, allowing warehouses to grow with the technology. Distribution centres that use this technology no longer need to allocate manpower for bringing products from storage to loading or sorting bays, and a series of instructions can be put in place to allow the workforce to get on with other tasks. Safety concerns can also be reduced, as automation takes employees out of harm’s way. The risk of colliding with people or other AGVs is low, as the decision-process has been removed. When danger does occur, the robots can simply stop, without any thought.

Imaging systems also offer a way of addressing the ‘bin-picking’ problem (see feature on page 20). Bonkenburg said: ‘The future is going to be, I believe, robotics within the warehouse, where it’s using a camera to locate and pick up objects.’ Within the warehouse, objects are often randomly arranged within a bin, making this is a problem of two parts, explains Bonkenburg: ‘The first being that it’s really hard to find out where one object is in among the jumble of everything else and say “OK that is really an object”, and the second is then “how do I actually pick it up?”. You have to solve both problems to do something really interesting.’

Some companies have addressed this already and use a combination of lasers and cameras to recognise shape, depth, and decide what objects are on top of others. This allows different types of objects to be picked from a bin and sorted. However, until now this has been most successful on a scale too small for use in a warehouse – but can be used to palletise and de-palletise products for storage or distribution, and to load or unload goods bearing trucks.

‘The more that you can use a robot to look, indentify, see and then pick something up the more you have a chance to automate processes that are not currently automated,’ Bonkenburg concluded.

Related features and analysis & opinion

03 January 2019

Warehouses are becoming highly automated facilities that rely to a large extent on vision, as Andrew Williams discovers

20 June 2019

The UK is up to 20 per cent less productive than its major competitor countries because it is not investing in automation, Mike Wilson at the British Automation and Robot Association said at UKIVA's machine vision conference in Milton Keynes. Greg Blackman reports

29 March 2019

Andrew Williams explores the vision solutions for robot bin picking

03 January 2019

Warehouses are becoming highly automated facilities that rely to a large extent on vision, as Andrew Williams discovers

15 November 2018

In an effort to improve operational efficiency and minimise or eradicate defects, a growing number of manufacturers in the automotive sector are introducing machine vision technology in quality control processes for components, modules, sub-assemblies and finished vehicles. So, what are the main current and potential applications of machine vision technology in automotive manufacturing quality control applications? What are likely to be the key innovations and trends in this area over the next few years? What role might vision technology play in the automotive factory of the future?