The Third Vision System is a holistic sorting system for packages that could be put to excellent use in sorting centres, warehouses, and any industrial environment where human labour is needed to sort goods.
Automated OCR (optical character recognition) processes are implemented to sort packages and letters in large-scale sorting centres for the express industry. Manual handling cannot be completely avoided however, especially for odd shape parcels that can’t be placed on the main conveyor. It takes three operators and twenty seconds to sort just one parcel, and manually keying the postcode and address creates the biggest bottle neck for the sorting process. Some problems associated with current hand-held scanners (that are used by the operators) are that they recognise only barcodes, and they are attached to fixed points by cables. Some state-of-the-art hand-held systems can perform OCR, but are still attached to a fixed point.
The Third Vision System emerges as a completely intuitive and wireless concept to support the operators. It uses gestural interaction. The system incorporates the third eye cam (a wearable head-held camera) and the third vision software (a program that uses pattern-recognition algorithms). The system recognises framing hand gestures in front of the camera and performs OCR directly from the parcels’ surface.
Third Eye Cam
The wireless head-held wearable camera gathers images of the parcels’ surfaces from the operators’ point of view. Hand gestures control the camera.
Special dot-patterned gloves are needed because bare hands doesn’t provide enough contrast for recognition by the optical censors. For this reason, the dot patterns were created in ski blue because this is a very unlikely colour for packages. (The colour can be changed to suit the user’s needs.) The camera was designed using widely used technologies that make it producible at reasonable costs: HD CMOS optical censors (wide angle), ultra-thin lithium batteries, OLED display, induction charging, Wi-Fi, LED projectors, and mono speakers.
Third Vision Software
The software uses pattern recognition algorithms to track the framing gestures. Such a gestural interface offers the best approach to controlling the camera and indicates to the system which image to take. The interaction creates an encouraging, motivational, and challenging activity based on hedonomics (or design for happiness). It offers the possibility of improving the operator’s performance while they actually have fun – a stress reduction strategy.
The framing gestures create square regions indicating to the system the area to cut out from the image – that is, the area that doesn’t contain characters to be recognised. This low-resolution image is sent by the third eye cam via Wi-Fi to the network, where the OCR engine decodes the character data and compares it with the sorting centre’s database. This approach removes the heavy OCR processing from the camera, lowering its power consumption, and consequently its weight and heat generation, achieving a comfortable and wearable device to for use by the operators.
Concept Test
A working model was built and connected to a computer where the preliminary version of the Third Vision software was running. Eight subjects were tested, achieving an average time of fifteen seconds while performing the framing actions over the parcels’ surfaces. Tests showed that using the Third Vision System, the sorting time can be reduced to seven seconds (best performer). With this new interaction, the operators’ moving actions are reduced, potentially improving the sorting task by almost 50%. |
