Digital Assistant for Tea-Estates - How IoT can help Tea-Estates reduce losses - Case Study

Flow of data from GreenSense nodes to Yuktix Gateway and to cloud.

Introduction

Apart from water, tea is the most popular beverage in the world. It is believed that tea originated in China and they are the largest producer of tea in the world followed by India and Kenya. Below is a map which show the distribution of tea production all across the world and you can clearly see the Asian sub-continent as the biggest producer.

Figure - 1 Top tea producers of Tea in World

And the top 10 countries growing tea are China, mainland, India, Kenya, Sri Lanka, Turkey, Vietnam, Indonesia, Iran, Myanmar, Argentina.

Figure - 2 Top 10 producers of Tea in World

Tea in India
Historical records indicate the prevalence of tea drinking in India since 751 BC. In the 16th century, a vegetable dish was also being prepared using tea leaves with garlic and oil. However, the credit for rediscovering tea and cultivating it at a commercial level goes to the British. It was Warren Hastings, Governor-General of Bengal who send tea seeds from china to Bhutan in 1774. You can refer this link to read more about the history of cultivation of tea in India.
In India, Tea is grown majorly in Assam, West-Bengal, Tamil Nadu, Kerala, Karnataka and some of north Indian states like Himanchal Pradesh, Uttarakhand, and Bihar. Below is a map which show the major states growing tea in India.

Figure - 3 - Top Tea producing states in India ( Source - Tea Board of India )

Interesting Fact about the name “Tea”

Tea is the only crop that is only know by 2 names universally - Cha and Tea based upon from which location it is imported. The one which came via sea is called Tea and the one which came via land is called cha. You can read this article to know more about this interesting fact.


The need of Digital Tools and Sensing in Tea- Estates
The demand for a pesticide free Tea is high after the Greenpeace report on Indian Tea [1][2]. Also, the consumer demand high quality tea and producers required high yield with low input cost. Above requirements can be achieved by

1. Elimination of shades of Trees
2. Reduction of losses due to disease and pests
3. Reduction to usage of pesticide
4. Weather Forecasting
5. Frequent soil test from multiple locations

With disease infecting the plants tea / coffee, estates managers are not left with no other option but to spray pesticides and insecticides throughout the estate. In agriculture, tea is the second crop to receive highest pesticide. This sometime leads to excess spray which leads to more investment in spraying, labor and deterioration of quality of tea.

Tea plants are attacked by several pests and diseases which are responsible for 7–10% loss in crop[3]. Recently in Kerala and Karnataka regions, due to heavy rain, there is outbreak of fungal diseases like black rot and stalk rot. There are many other disease which affects the tea plants, some of them are following.

1. Blister Blight 
2. Red Mite 
3. Tea Mosquito 
4. Black Blight, thread blight
5. Grey Blight 
6. Red Rust
7. Pink Disease 
8. Hypoxylon wood rot

Digital tools and Smart sensing in tea-estate is required to reduce the

1. Usage of pesticide and insecticide 
2. Loss of attack due to pest and disease 
3. Frequent degrading of soil 
4. Residue in Tea
5. Manual error in recording incidences 
6. Wastage of man-power in sharing data across the operations and management team

Client

One of the biggest tea-producer in South India


Product

Yuktix Digital Assistant for Tea-Estate and risk management

Problem - Most of the tea-estates are situated in remote locations with limited connectivity. As mentioned above tea is the second crop that receives the maximum pesticide.

1. Usually in tea-estates the identification of disease is done by manual process i.e. by visiting each hill and collecting the leave samples which are analyzed later. This manual process is repeated every week without backed by data.
2. The spray of pesticide is totally based on guess work (almost 3 times in a month). After spraying pesticide it take need almost 14 days before you can pluck the tea-leaves.
3. Spray of manure is based on soil moisture which right now is not measured. 
4. The current mode of  capturing weather data is very manual and centric i.e. it is collected in the tea-factory whereas there can be a variation of almost 10-25% within a distance of 1 km in a tea-estate. Right now there is no way to automate the process of capturing granular weather data.

Challenges

1. Remote off grid, hilly terrain with limited network connectivity 
2. Harsh Environment 
3. Cloudy environment with sometime no-cloud for a week and continuous rainfall.  
4. Tea-estate are spread across 100's of hectare 
5. Forest area with elephant crossings. Not safe after sunset.

Solution
Yuktix Digital Assistant for Tea-Estate monitoring and Risk Management

Yuktix GreenSense in action in a tea-estate

Features of Yuktix - Digital Assistant for Tea-Estates

• Battery powered GreenSense Devices with Air Temperature, Air Humidity and Soil VWC sensors 

Yuktix GreenSense with Soil VWC

• Solar powered GPRS / Low earth orbit gateway 
• Solar powered Automatic Weather Station with Rain Gauge and Leaf Wetness Sensor.

Yuktix Weather Station in a Tea-Estate

And GreenSense - Tea dashboard with actionable insights like 

1. Weather Forecast 
2. Sensor data 
3. Disease and Pest prediction 
4. Soil Health card 
5. NDVI and EVI images for hot-spot detection 
6. Image analysis
7. Alerts, reports and Farm activity recorder. 

Yuktix NDVI Image of tea-estate covering 250 acres

How does it work?

The target micro-weather monitoring locations are identified. Yuktix battery powered wireless GreenSense nodes are deployed in these identified locations. All the GreenSense nodes communicate wireless back to Yuktix Solar powered gateway. Data collected by the gateway is then pushed to Yuktix cloud via GPRS or Low earth satellite communication. 

Flow of data from GreenSense nodes to Yuktix Gateway and to cloud.

The collected data is received by ankiDB IoT cloud, where is processed and stored in a time-series database. The collected data is then tunneled in real time to various instances of Yuktix compute engine running different processes like Red mite disease machine learning, tea blister blight algorithms, weather forecast, alert rules, and email reports. The process data is then entered into a message queue to process further and made available to Yuktix GreenSense dashboard via REST API's. 

Impact  

With Yuktix Digital Assistant for Tea-estates, the operations team was able to

1. Reduce the pesticide spray by 15% which amount to USD 500 a month. 
2. Reduce the man-work hour spend in scouting by 150 hours a month which amount to 200 USD a month. 
3. Able to reduce the residue on the tea-leafs by 60-70%. 
4. Improve the data capturing process. Now the data is much organized and available at a click to operations team and management team. 
5. Operations team is able to optimize the resource usage by 50% as compared to earlier.