Beyond Binoculars: AI Bird Counting and the Future of OrnithologyPosted: April 2, 2023 Filed under: Uncategorized Leave a comment
Artificial Intelligence (AI) bird counting is a rapidly growing field that is revolutionizing the way we monitor bird populations. Bird counting is an essential tool for wildlife conservation and management, as it provides vital information about population trends, habitat preferences, and migration patterns. AI bird counting has the potential to make bird counting more accurate, efficient, and cost-effective. Let’s talk about the application, limitation and recently development of AI bird counting in Hong Kong.
AI bird counting has several applications in the field of wildlife conservation and management. One of the significant benefits of AI bird counting is its ability to detect and identify bird species accurately and quickly. This technology can provide real-time data about bird populations, which can be used to monitor and manage bird habitats. Furthermore, AI bird counting can automate the bird counting process, which can save time and resources.
Point count is a common method used by ornithologists to estimate bird populations. This technique involves observing birds at a fixed point for a set amount of time. However, this method is time-consuming and often needs more than one experienced observer. Especially, in the migration bottlenecks where bird migrations occur intensively in specific periods, the observer has to detect and count birds and keep records. Furthermore, the success of point counts is much obliged to the human census and affected by the observer’s ability and experience, environmental and topographic variables and birds’ detectability. Detection of birds in bad weather conditions such as intense cloud cover, rainy and foggy weather, and low light intensity might not be possible. As such, researchers are continually developing new techniques and technologies, such as AI bird counting, to improve the accuracy and efficiency of bird population monitoring.
Despite the potential benefits of AI bird counting, there are also some limitations to this technology. One of the significant limitations is its accuracy in detecting and identifying bird species in complex environments. AI bird counting systems may have difficulty distinguishing between similar-looking bird species, which can lead to errors in population estimates. Additionally, AI bird counting systems may not be effective in detecting birds that are hidden or camouflaged, such as birds in dense vegetation.
Major misdetections occurr due to very small bird sizes. As the distance to birds increases, such birds often correspond to very small details (i.e., several pixels) for even the human eye to differentiate. Small object detection is also an open-problem in many other datasets. Other main reasons for the misdetections are intra-species variation, unusual bird poses, occlusions by other birds and plants, cast shadows, and background clutter.
AI bird counting technology is rapidly evolving, and researchers are continually working to improve its accuracy and efficiency. One of the significant developments in AI bird counting is the use of machine learning algorithms, which can learn from large datasets and improve their accuracy over time. RoboticsCats partners with Hong Kong Bird Watching Society (HKBWS), World Wide Fund for Nature Hong Kong (WWF-Hong Kong) and APlus Digital Academy (APlus) to explore AI bird detection and counting in Hong Kong. We deployed fixed and mobile cameras in multiple locations to evaluate AI bird detection performance.
We find the AI detection confidence score will significantly affect the bird counting performance. If we require AI detection with higher confidence score, there will be more false negatives and the system will underestimate the bird population. Vice verse, if we accept AI detection with lower confidence score, there will be more false positives and the software will overestimate the number of bird. We are looking for the most effective confidence score values in our future researches.
Additionally, we are developing new sensors that can capture high-quality data in challenging environments. Another development is the use of citizen science initiatives to gather data for AI bird counting. Citizen scientists can take photos of birds and upload them to databases, which can then be used to train AI algorithms. This not only helps to improve the accuracy of the algorithms but also engages the public in conservation efforts.
In conclusion, AI bird counting is a promising technology that has the potential to transform the way we monitor and manage bird populations. While there are limitations to this technology, ongoing developments in AI algorithms and sensing technology are likely to improve its accuracy and effectiveness in the future. As such, AI bird counting is a field to watch for conservationists and researchers who are committed to protecting the world’s bird species.
- Automated Bird Counting with Deep Learning for Regional Bird Distribution Mapping
- AI Wildfire Detection Accuracy: Precision, Recall, mAP
Black-faced Spoonbills in Hong KongPosted: March 14, 2023 Filed under: AI, CDN | Tags: biodiversity, video streaming Leave a comment
A photo tells a thousand words.
The AI bird detection system in Mai Po Nature Reserve captured some good photos of the Black-faced Spoonbills last week.
The Black-faced Spoonbill (Platalea minor) is a medium-sized wading bird that is named for its distinctive black facial skin, which contrasts sharply with its white plumage. It is one of six species of spoonbills, and is currently listed as endangered by the International Union for Conservation of Nature (IUCN). There are just 6,162 Black-faced Spoonbillls in the world.
Black-faced Spoonbills are found primarily in East Asia, including China, Japan, South Korea, and Taiwan, as well as in parts of Southeast Asia. They typically inhabit coastal areas, estuaries, mudflats, and other shallow wetlands, where they feed on small fish, crustaceans, and other aquatic invertebrates.
The Black-faced Spoonbill was once hunted extensively for its feathers, which were prized for use in hats and other fashion accessories. Habitat destruction, pollution, and other human activities have also contributed to the bird’s decline in recent decades. Conservation efforts, including habitat restoration and protected areas, have helped to stabilize the Black-faced Spoonbill population, but ongoing threats mean that the species remains at risk of extinction.
The Mai Po Nature Reserve in Hong Kong is an important habitat for the Black-faced Spoonbills.
There are several ways in which we can help to protect the Black-faced Spoonbill:
1. Habitat conservation: The Black-faced Spoonbill’s natural habitat, such as coastal wetlands, mudflats, and estuaries, should be protected and restored. This can be done by designating protected areas, enforcing regulations against habitat destruction, and promoting sustainable land use practices.
2. Pollution control: Pollution can affect the quality of the Black-faced Spoonbill’s habitat and food sources. Measures should be taken to reduce pollution from industrial and agricultural sources, as well as from plastic waste in the oceans.
3. Sustainable fishing practices: Overfishing can deplete the Black-faced Spoonbill’s food sources. Sustainable fishing practices should be promoted to ensure that the bird’s food supply is not threatened.
4. Public awareness: Raising public awareness about the Black-faced Spoonbill and its conservation status can help to generate support for conservation efforts. This can be done through education campaigns, bird watching tours, and other outreach programs.
5. Research: Research into the ecology and behavior of the Black-faced Spoonbill can help us to better understand its needs and develop more effective conservation strategies.
By taking these steps, we can help to protect the Black-faced Spoonbill and ensure that this beautiful bird does not go extinct.
Want more photos? See this instagram.com/roboticscats/
From Wildfire to Wildlife: The Curious Engineer Who Developed Real-Time AI Bird Detection SystemPosted: February 26, 2023 Filed under: Uncategorized Leave a comment
Once upon a time, there was a curious engineer named Alan who was fascinated by the power of artificial intelligence. Alan was always looking for ways to use technology to solve real-world problems. One day, while reviewing images from a wildfire detection system, he noticed a Turkey Vulture flying through the frame. However, he had missed the bird during his initial review and became curious about how he could develop a real-time AI bird detection system using existing surveillance cameras.
Alan began his research by studying AI bird detection models in the markets and their applications. He then developed a prototype that could detect birds in real-time using existing surveillance cameras. Alan’s prototype used machine learning to analyze live video feeds from the cameras and identify birds based on their size, shape, and posture. The AI model also could distinguish between different bird species such as Great Egret and Great Cormorant.
The system proved to be a success during proof-of-concept and Alan’s invention was widely praised. The AI bird detection prototype was particularly useful for tracking birds near wetlands, which are important habitats of migratory birds. Alan’s invention was also beneficial for wildlife conservation efforts, as it enabled researchers to study bird behavior in their natural habitats without disturbing them.
Alan’s invention inspired other engineers and researchers to explore the potential of AI for wildlife conservation. Alan’s curiosity and passion for technology led him to develop a cost-effective, quick-to-deploy, and easy-to-use AI bird detection system that had the potential to make a significant impact on wildlife conservation. His invention served as a reminder that technology can be used for good, and that with a little bit of creativity and determination, engineers can develop solutions to even the most unexpected problems.
(This story is originally written by ChatGPT.)
RoboticsCats is among the Top 50 startups for the 4YFN Awards 2023Posted: November 7, 2022 Filed under: AI, CDN, mobile, startup, wildfire detection | Tags: AI, CDN, mobile, startup Leave a comment
The 4YFN Awards is a global startup competition aimed at finding the best digital startups around the globe. It is the highest official recognition offered exclusively to startups during MWC Barcelona 2023.
This year, 4YFN received hundreds of startup applications from no less than 49 countries around the world, an all-times record for the organisation. Our startup RoboticsCats is thrilled to have been selected as part of the exclusive Top 50 startups list, reserved for companies who ranked at the top on criteria including innovation, soundness, traction, potential, impact and fit.
Humanising Technology, 4YFN 2023 event theme, speaks about Startups as a force of disruption. The event will showcase and debate how entrepreneurs and their ecosystems can build a more humane technological playing field to solve society’s current and upcoming challenges. We are 100% on board with this mission and are pleased to have the chance to form part of this journey.
On Friday Nov 11, 2022, 4YFN will be announcing the 5 finalists of the competition, chosen by a panel of industry experts and investors from organisations like Playfair Capital, LUMO labs, Draper B1, Harbinger VC, GSMA Foundry, Seaya and others. Stay tuned to the announcements here and on 4YFN social media channels
RoboticsCats is a digital and mobile startup. Our first product is the Insight Globe Mobile (IGM) app launched in February 2020, six months after we founded the company. IGM empowers InsightFD wildfire detection system customers to easily receive and use image-based wildfire detection alerts on their smartphones
The COVID-19 outbreak accelerated our mobile-first strategy. We launched our consumer product ReportFires web app in March and the ReportFires app on Kickstarter in October 2020 respectively. ReportFires is a free crowdsourcing platform to facilitate people to report and receive wildfire outbreak locations. Modern map app design, single hand operation which takes less than three seconds to report and share wildfire location.
We built and installed the world’s first AI wildfire detection service-as-as-service deployment at Big Tree Animal Sanctuary and Adoption Centre in March 2021. We connect AXIS Communications HD and 4K surveillance cameras to a congestion-free 5G network via a Netgear Nighthawk M5 router. The animal center users use ReportFires app to receive detection alerts. The design and lessons learned in serving this non-profit customer became the foundation of our LookOut wildfire detection SaaS.
Our project was nominated by the Hong Kong Government as one of the innovations funded by the Subsidy Scheme for Encouraging Early Deployment of 5G.
Between June and September 2021, we implemented our first and second international LookOut customers at Piriapolis in Uruguay and Barra do Garcas in Brazil. They are innovators. We shared their success stories at the UNEP Modern Technologies for Disaster Management Webinar on 28th September 2021. Since then we have been working with customers to use LookOut service to protect forests, communities, and infrastructure in Asia, Europe, and the Americas.
We focus on global partnerships in 2022 to innovate faster and better. We team up with startup communities, research institutes, and NGOs to add new capabilities to the LookOut platform. One new tool is our LookOut Cam app which turns an Android phone into an image input device for our LookOut wildfire detection SaaS. We shared our mobile strategy at ViennaUp’22 Connect Day and demonstrated the prototype at the 5G International Exchange Forum in Vienna.
Top 50 startups for the 4YFN23 Awards is a strong encouragement to our digital and mobile-first strategies over the last few years! Kudos to our customers and partners. We work together to build the world’s most mobile-friendly wildfire detection solutions. It is a long and rewarding journey: Hong Kong, Korea, Vietnam, Thailand, Indonesia, USA, Brazil, Uruguay, Chile, Mozambique, Latvia, Greece, Portugal, and Spain.
Our LookOut Cam app, LookOut wildfire detection SaaS, and ReportFires app provide an end-to-end mobile AI wildfire detection solution. It empowers users to use our technologies at their fingertips. All we need is an internet-connected smartphone.
Register your free LookOut wildfire detection SaaS trial account and download LookOut Cam app now.
What if?Posted: June 18, 2022 Filed under: Uncategorized Leave a comment
Time flies. Today is my 7th day back in Hong Kong after the two-week Austria business trip. It is also my 7th and last day in the quarantine hotel.
Other than business-related online meetings and phone calls, I don’t have any real-time interaction with other people. It is a moment of solitude.
Four Points by Sheraton Hong Kong Tung Chung is a nice hotel. Clean and tidy. I can see green mountains, blue sea and sky from my room. I can sometimes see cars, ships, UFO (unidentified floating objects), cable cars, birds, and an imaginary wildfire…
Solitude is creativity’s best friend.
What if there is a wildfire on the opposite mountain at midnight when I am sleeping? What if between the opposite mountains and me is a dry grassland? What if a strong wind blowing in my direction spreads the wildfire quickly to the hotel? What if I am not allowed to leave the quarantine hotel even when it is burning?
What if I point my smartphone to the mountains and use it as a camera for the LookOut wildfire detection service? What if LookOut detects the initial wildfire outbreak, sends alerts to the fire department to help them take timely action to control the fire, and saves me and my neighbors? What if I create this solution?
A long and silent Saturday afternoon.
Wish everyone a joyful Sunday tomorrow.
My demo day in Latvia, 31-March-2022Posted: April 10, 2022 Filed under: Uncategorized Leave a comment
I presented on the Silvatech Public Demo Day on 31-March-2022. Some of the cohorts presented at the event venue in Latvia while some others participated virtually. I joined remotely from Hong Kong.
On top of topics related to forest tech, the Silvatech Growth Accelerator program covered areas that are common in a well-structured 12-week accelerator program – team building, product development, sales, marketing, investor pitch, lead generation, market engineering, competitive analysis and fundraising, etc. A busy schedule.
As a first-time entrepreneur, I found the program was very practical and productive, especially those two-week workshops focusing on Product Sprint and Market Engineering. I learned a lot by experiencing and observing how much we and other cohorts traveled and improved by working thru proven methodologies.
On the Demo Day, there was a short introductory animation for each startup before its presentation. The animation thing was a very nice surprise to me. It was a bigger surprise to me when I watched the animation for Robotics Cats:
I recalled in the application interview last December, LVM and SWG interviewers asked me why Robotics Cats is interested to join this Latvia-based accelerator program. “First of all, I am very keen to learn more about forest tech from the program and other cohorts. Secondly, the program is mostly virtual, I can and will be committed to participating in it. Last but not least, it was not our first time in Latvia. We did join our partner to take part in the wildfire detection competition in Q1 2020. We were not well prepared and did not win last time. It was graceful that our partner supported us to do a pilot after the competition to improve our AI wildfire detection. We improved a lot in the last two years with more pilots and customers in Europe. We are ready now and so we come back.” I answered them and showed them the following slide.
It was so nice when I saw the animation. They listened to us. I appreciate it a lot. Thank you very much.
The 2021-22 SlivaTech program is completed successfully. All cohorts graduated. Bravo!
I join the community formed by entrepreneurs who previously attended other startup programs organized by Startup Wise Guys.
I am now a Startup Wise Guys alumni. It is so cool!
UNEP Modern Technologies for Disaster Management Webinar SeriesPosted: October 3, 2021 Filed under: Uncategorized Leave a comment
This is a long weekend in Hong Kong. I have some time to write something. Thanks.
It was my pleasure to present in the UNEP Modern Technologies for Disaster Management webinar last Tuesday (28-Sep-2021). This webinar focused on opportunities and challenges associated with Artificial Intelligence and its use in disaster management. It was the second webinar of the four-part series. The first one (14-Sep) was about Robotics, the third (12-Oct) and fourth (26-Oct) webinars will talk about Drones and the Internet of Things. Please register at: bit.ly/ModernTech4DisasterMgmt
It is a new exposure for me to speak in a UNEP event There are some things very interesting and I would like to share with you.
There were five speakers and six members in the UNEP organizing team (Resilience to Disasters and Conflicts Global Support Branch). These 11 people are based in 11 different countries: Germany, Qatar, China, France, United States, Switzerland, Netherlands, Botswana, United Kingdom, Nepal, and Mexico. I found the teamwork was very good and professional. Kudos to the organizing team!
The webinar attendees asked very good questions! Domain-specific and relevant. Very impressive.
The speakers were AI experts! I learned a lot from this event! It was a very informative webinar. In case you missed the webinar, I highly recommend you watch the video recording. I think you can find it on YouTube.
If you want a quick summary of the presentations, below are my picks.
Opening remarks, by Monique Kuglitsch:
“High-quality data are the foundation of AI-based algorithms.”
It is so true and fundamental!
Dear customer, please clean your camera lens and provide us clear images. Thanks.
Extracting information from social media for crisis response and management, by Dr. Muhammad Imran:
“Deep learning models need, first of all, huge amounts of data. Second, they need to have diversity. So having the footprint of one incident in different cases would greatly help deep learning models to make more robust inferences.”
AI-powered wildfire detection system, by me:
“We need different tools for different stakeholders to stop the threat from spreading. Masks for COVID-19. Early detections for wildfire.”
Monitoring and forecasting floods with artificial intelligence, by William Castaings:
“Velocity and discharge remain quantities of interest which are very often challenging or impossible to measure during extreme events.”
Integrating AI with hazard science to improve disaster resilience, by Craig Fugate:
“This is really about how do you price risk in a way that induces both governments, but more importantly private sector, to change behaviors both in reducing carbon emissions and other greenhouse gases, but also actions they need to take now to mitigate against the impacts.”
“By working together with the right tools, we can bring positive impacts on a global scale today.”
This UNEP webinar series is one the right tools. We need more. I am looking forward to the coming Drones and IoTs webinars! Please join us and register at: bit.ly/ModernTech4DisasterMgmt
RoboticsCats in AcceliCITYPosted: June 16, 2021 Filed under: Uncategorized Leave a comment
It is an exciting journey.
There are 50 semi-finalists in this year’s QBE AcceliCITY Resilience Challenge program. We are excited that RoboticsCats is one of the 50 startups!
What is AcceliCITY?
AcceliCITY is one of the largest Smart and Resilient City startup accelerators in the world. It focuses on startups that utilize smart city solutions to address risk, equity and sustainability in our urban environments.
Who is the program organizer?
Leading Cities, an international non-profit organization, connects smart cities across the globe with innovations and insight to drive resiliency, equity, and sustainability. It is headquartered in Boston, Massachusetts, USA.
What is the accelerator program content?
The 5-month accelerator program is designed to help startups navigate the challenges of working with governments (YEAH!), convey venture value, and network within an expansive Smart and Resilient City ecosystem.
Because of COVID-19, the 2021 program will mainly consist of webinars and online workshops. Usually, the program will start at 10 am Eastern Time (10 pm Hong Kong Time).
What are the monetary benefits to the accelerator cohorts?
A US$150,000 pilot project!
The program culminates with a final pitch event where up to $150,000 in city pilot opportunities are awarded.
AcceliCITY’s direction is to help startups to do business with cities (governments) and AccelCITY will accelerate the selling cycle by funding a pilot program. For companies targeting government customers, the first pilot deployment is very critical! We know that! And that’s why we’re keen to participate in the program!
“At Leading Cities, we recognize that startup growth and impact require capital from both investors and clients. The AcceliCITY program connects winners with funding, a pilot project, and a pipeline of clients from cities worldwide.”
Who are the 2021 AcceliCITY Semi-finalists?
There are 500+ applicants from 44 countries. The 50 semi-finalists are from 23 countries:
20.South Korea: 1
The 2021 AcceliCITY accelerator program has just started. Many new and interesting things. I gave my first Clubhouse speech yesterday! Cool!
It is our honor to join the other 49 startups in this exciting journey!
Big Tree protects Cats and Dogs with LovePosted: June 5, 2021 Filed under: Uncategorized Leave a comment
“Big Tree Animal Sanctuary and Adoption Center (hereafter Big Tree) is a charity organization in Hong Kong. The center is located in a remote rural area surrounded by hills and trees. It is facing high wildfire risks, especially in April and October. Big Tree uses 5G and RoboticsCats AI-Cloud SaaS to early detect wildfire and mitigates wildfire risks to the center and the animals living in it.”
Figure 1: a camera installed at Big Tree Animal Center. The center is located at the foothill.
Above is an extract from the company blog post with the title “Big Tree protects Cats and Dogs by 5G & AI” that I wrote three months ago. It was a good article. That said, with the new experience and lesson learned since then, I should rewrite the title, right the wrong.
Big Tree protects Cats and Dogs with Love.
Big Tree is the first animal center in the world to use AI to detect wildfire. It is an innovative application. Many media  are interested to report how Big Tree uses technologies to mitigate wildfire risks. I joined some of the media interviews. The reporters asked very detailed questions. The media crews worked long hours with smiles. Their narratives and videos are both informative and beautiful.
Figure 2: three different media interviewed Big Tree on the same day!
It is joy. It is love.
All the cats and dogs living at Big Tree were abandoned in the past because of their physical sicknesses, disabilities, or aging. The abandonments created more mental damages and broken hearts which are far more challenging to recover than physical wounds. The animals felt remorse, sadness, distress and loss.
Big Tree protects the cats and dogs with love.
Figure 3: Reporter and photographer were busy working on the interview.
In one of the interviews, a reporter asked Big Tree volunteers to share their experience with the new technology. “The new tool helps us to early detect nearby wildfire, and we can evacuate the animals timely if needed and save their lives, it probably can save the animal center structure too. Even better is, we believe the early detection will save the lives and properties of our neighbours further away from the foothill.” Big Tree volunteers replied.
A surprising and inspiring answer.
Figure 4: Neighbours of Big Tree: farms and residential properties
“And a second is like it: You shall love your neighbour as yourself.” Matthew 22:39 ESV
Figure 5: Neighbours of Big Tree from a different perspective
Please contact Big Tree at its website or Facebook page if you want to support them.
 Below are the recent media reports and case studies about the Big Tree project, in reverse chronological order:
5G x AI 偵測山火 Case Study: NETGEAR M5 5G Router 與 RoboticsCats 保護大樹下庇護站貓狗
Netgear SMB, 26-May-2021
Axis Communications and Robotics Cats provide early wildfire detection to animal sanctuary
A&S Magazines, 17-May-2021
HK Cable TV, 15-May-2021
Customer Stories: Early wildfire detection for Big Tree Animal Sanctuary and Adoption Center
Axis Communications, 15-May-2021
Axis Communications 夥Robotics Cats 為大樹下提供山火預報解決方案
IT Pro Magazine, 13-May-2021
“Big Tree” project (Axis Communications, Robotics Cats & Armanda)
Armada International, 6-May-2021
Apple Daily, 18-Apr-2021
StartUpBeat, Hong Kong Economic Journal, 22-Mar-2021
Netgear M5 5G router review in Hong KongPosted: January 24, 2021 Filed under: Uncategorized Leave a comment
I tested Netgear Nighthawk M5 5G WiFi 6 Mobile Router last weekend.
1.Why are we interested in 5G?
With the rollout of 5G, cellular data communication services have become more accessible and affordable. We are exploring cellular services (5G, LTE-A, LTE) to connect our wildfire detection hardware.
Most of our deployments are using fixed microwave wireless networks to connect our robots in forests. To move to public 5G/4G cellular services will lower the network management cost and improve cash flow.
An important challenge is cellular signal strength in remote/rural locations where most of our robots/cameras are installed.
We also need a good router that can provide reliable cellular uplink, and wired Ethernet ports to connect to our robots/cameras. We want small physical sizes and power consumption.
2.Netgear Nighthawk M5 5G WiFi 6 Mobile Router
There are only a few 5G routers in the Hong Kong market. Netgear M5 is one of them. The one we are testing is an unlocked version of Netgear MR5100. It is effectively an MR5200.
Netgear M5 supports device connections via WiFi, USB-C tethering (for both Windows PC and Mac), wired Gigabit Ethernet (requires Data Offloading turn-off). It can be powered directly via USB Type-C with the internal battery removed. It is designed as a mobile router, hence it is small (105 mm x 105 mm x 21.5 mm)and light (240g).
I tested its uplink performance in the rural area. We will use Speedtest app and iPhone 11 Pro which connects to M5 via WiFi. The testing location was Cape Collinson Battery. It is located in eastern Hong Kong Island. 3.5km walking distance from the Chai Wan MTR station. I tested three times, fastest download and upload speeds at 83.8Mbps and 23.2M respectively! Very good!
3.More about Cape Collinson Battery
Two batteries are built at Cape Collinson in 1938. We can see only one of them now. It becomes a popular eco-travel spot because of its unique history and beautiful landscape. To test the Netgear M5 at Cape Collinson Battery makes the whole journey more interesting!
“I’m going to live every minute of it.” Joe Gardner