Seeing the unseen: How Mitti Labs uses remote sensing & GIS to decode agriculture

At Mitti Labs, we’re building smarter tools to understand agriculture at scale without needing to set foot in every field. Thanks to remote sensing and Geographic Information Systems (GIS), we can analyze crops, soil, water, and field conditions across entire regions using data captured from space, aircraft, and drones.

We capture insights from a variety of sources, from drones for ultra-high-resolution studies to satellites that let us monitor entire states or countries. This blend of scale and detail is key to delivering scalable, hyperlocal insights.

In this blog post, we dig into how we apply this technology to transform raw satellite signals into actionable insights for farmers, researchers, and carbon credit buyers.

Remote Sensing: The engine behind our insights

Remote sensing, at its core, is about capturing information without physical contact. Think of it as a high-tech extension of our eyes and ears. Remote sensing systems can be passive (relying on sunlight, like traditional satellite imagery) or active (generating their own signal, like radar). At Mitti Labs, we use both.

Passive sensors help us map vegetation using visible and infrared light. Active sensors—like Synthetic Aperture Radar (SAR)—allow us to "see" even through clouds or at night, which is crucial for places with frequent rain or persistent cloud cover.

Making sense of the spectrum

One of the secrets behind our tech is how we use the electromagnetic spectrum. Different wavelengths (like visible light, infrared, or microwave) interact differently with surfaces like crops, soil, and water.

For example, we use near-infrared bands to assess crop health; healthy plants reflect more NIR than stressed ones. Our systems also tap into microwave data to monitor soil moisture and detect flooded areas, even when optical sensors fail to see the vegetation because of a large cloud cover.

We classify this into optical, thermal, and microwave remote sensing. Each brings its own strengths. Optical tells us what’s happening on the surface. Thermal detects emitted heat, helping us identify irrigation patterns or crop stress. Microwave, however, is where it gets really exciting.

Why we love microwave sensing

At Mitti Labs, we’re big believers in microwave remote sensing, especially L-band radar. Unlike visible light, microwave signals can penetrate clouds, vegetation, and even shallow soil layers. That means we can monitor fields in India’s monsoon season or during early sowing periods when seeds are still covered by soil.

Different microwave bands (X, C, L) offer different levels of penetration. L-band is great for looking into soil moisture and root-zone conditions. C-band gives us a canopy structure and biomass. X-band, with shorter wavelengths, primarily interacts with the upper canopy and surface features, making it useful for detecting fine-scale surface roughness and leaf-level characteristics. And with polarization techniques, we can detect whether radar waves are bouncing off vertical structures like crop stalks or flat surfaces like standing water. There is science behind this "see-through" ability: if the wavelength is longer than the target’s size, it passes through. That’s how L-band radar gets under the canopy while X-band bounces off the leaves.

It’s not just data—it’s resolution that matters

In remote sensing, not all images are created equal. We carefully select sensors based on four types of resolution:

• Spatial resolution: How detailed is the image, does one pixel represent 10 meters or 30 meters? This determines how small of a feature we can detect. For example, high spatial resolution helps us map narrow field boundaries, detect patchy crop growth, or identify small water bodies. It’s critical when precision and localization matter—like identifying which fields are waterlogged after a storm.
  
  

• Spectral resolution: How many colors or bands does it capture? This is key to identifying vegetation types. This impacts how well we can differentiate between crops, soils, or vegetation states. For instance, being able to distinguish wheat from rice or healthy plants from stressed ones depends on capturing reflectance across multiple, specific wavelengths (like red, near-infrared, and shortwave infrared).
  
  

• Radiometric resolution: How precisely can we measure energy differences? Reflectance refers to the proportion of light or electromagnetic energy that a surface bounces back after being hit by a source like the sun or a radar signal. Different surfaces—like healthy crops, bare soil, or water—reflect energy differently depending on their texture, moisture, and color. The higher the radiometric resolution, the more detail we capture in those differences.
  
  

• Temporal resolution: How often can we get new images of the same location? Is it daily or weekly? This is crucial for monitoring change over time—whether we’re tracking sowing dates, crop growth, flooding events, or harvesting activity. For example, high temporal resolution allows us to identify if a farmer missed a planting window or detect a sudden drop in vegetation index that signals crop damage.
  
  Our crop monitoring products require a careful balance—frequent revisits with enough detail to detect sowing, stress, or harvest trends.

Remote sensing meets GIS: The real power

Remote sensing gives us the raw inputs. GIS (Geographic Information Systems) is how we make them useful. We use vector data (points, lines, and polygons) to represent roads, field boundaries, and water bodies, alongside raster data (pixel-based satellite imagery), which gives us the actual surface observations. We combine these in our proprietary analysis pipelines to build products like:

• Crop calendars – Using NDVI and microwave data to estimate rice intensity in one particular area, sowing dates, growth duration, crop rotation and harvest windows.
  
  

• Field condition analysis – Identifying drought stress, saturation, or flood damage from radar backscatter and vegetation indices.
  
  

• Practice inference – Predicting whether fields are transplanted or direct-seeded using temporal profiles of vegetation and surface moisture.
  
  

All this happens within our custom geospatial architecture, built using tools like QGIS, open-source libraries, and cloud-native infrastructure for scalable processing.

Real-world application: From pixels to policy

Our work isn’t just academic. Our work is powering real-life decisions on the ground. At Mitti Labs, we’re helping agri-tech platforms, sustainability teams, and food system innovators embed climate-smart intelligence into their products and workflows. We're also building tools that support carbon credit verification in agriculture—tracking indicators like cover crop presence, residue retention, or biomass buildup over time. With high-resolution, remotely sensed data, we can validate whether carbon-sequestering practices are truly being implemented, enabling transparent and scalable access to emerging carbon markets.

 We’re helping:

• Prioritize irrigation investments

• Monitor compliance with climate-smart agriculture programs

• Deliver early warnings for drought or flood risk

• Optimize procurement and subsidy distribution based on real field conditions

We’re also supporting partners who are building farmer-facing applications by providing high-resolution, location-specific data that makes sense even without a satellite science degree.

Why this matters

At Mitti Labs we’re redefining how it’s applied to agriculture in the Global South. Our tech stack leverages cutting-edge earth observation, microwave sensing, and GIS to bring clarity to complex, fragmented agricultural systems. We believe that insights should be scalable, accurate, and actionable, no matter where the farmer is or what crops they grow.