"MINI-EDITH-THE SHEILD FOR WILDLIFE?"
Past few years has been a total misery for the living species (specially wild) to live on earth on ease of their own comfort & freedom. Many animals are limited by the number (in the verge of extinction); representing their own kind because of unwise selfish acts human are following these days like: poaching, trespassing the inhabitations, deforestation, violating animal rights, etc. However, acknowledging the significance that all animals have in co-existence, to balance our existing ecosystem and controlling animal's devastating behaviors; many digital brainstorming steps are inherited sequentially, out of many is Intel-powered AI Tech-Movidius Myraid Visual Processing Unit(VPU) camera: same used inside Google's Chip, Flair's Firefly, DJI's Phantom 4 cameras etc. are manufactured respectively.
Anti-poaching prototype, branded name- "TrailGuard" AI is a tool manufactured in April, 2018 under the bright sight of RESOLVE's WildTech Program. It was funded by popular actor's Leonardo DiCaprio Foundation, and National Geographic Society and whole system was setup & concepted by Anna Bethke, Head of AI for Social Good at Intel. The prototype is used as security system for National Park to detect, stop and arrest poachers in real-time and a survey system to track, monitor and better understand the ecosystem's health and biodiversity of wildlife.
CONSTRUCTION:
=> VPU packs dedicated hardware chips for computer vision and a dozen purpose built (SHAVE) processor cores to speedup execution of algorithm.
=> Size of a pencil roughly.
=> Front-View
=>Rear-View
Working And Features:
It uses Intel Movidius VPU for image processing, running deep neural network algorithm for object detection and image classification inside camera. Deep Neural Network Algorithm is used to selectively transmit photos over Local Network. Image Processing concept is taught to it to recognize suspicious act using thousands of photos taken from varieties of angles, poses and contexts, imbuing it with ability to make judgements about contents of the shots. If objects(vehicles & humans) are detected in any captured frames, it alerts park headquarters intelligently then the rangers team respond the false positives, primary attribute of TrailGuard AI. But for that it requires connected Microcomputers for data processing. It chooses images of abt 20KB to save bandwidth. It initially sent images over 3G GSM cellular which is of course a limited range of network. So, Developer Teams later installed cigar-box-sized satellite transmitters that communicate via LoRa (Long Range Wireless tech). It usually works under a very less power(nearly a Watt) in average day. It is designed to perform in the wild for up to 1.5 years without depleting the battery. For energy efficiency, the system uses passive sensor; only powers on when the sensor detects movement around its surrounding frame.
=>Sample of TrailGuard Detecting Its Object
------------------------------------------------------------------------------------------------------------------
' AI for Wildlife' Testing Process at Maasai Mara National Reserve section, Kenya
Field Tested:
Its first test was carried on at Grumeti Reserve in Tanzania which as a result reports arrest of 30 poachers and seizure of over 1300lbs. of illegal bushmeat in test phase. Different wildlife reserves and communities of Asia, South America & Africa are already using this camera tech starting with Serengeti and Garamba.
'AI Tested Successfully!' at Maasai Mara National Reserve, Kenya
0 Comments