Thermal image analysis for leak detection: settings, patterns and temperature difference

What is the Thermal Image Analyzer and what does it give you?

It is a tool inside the Kashf Pro app that analyses a photo or screenshot from your thermal camera, whatever the make, because it reads the colours of the palette itself.

It gives you the pattern of the largest anomalous area, its size, its direction and the temperature difference, plus an optional AI opinion that reads the temperature scale and compares before and after, then a report link you can share with the customer.

Who is this tool for?

• Leak detection technicians with a standalone thermal camera or one that clips onto a phone.
• Maintenance and waterproofing technicians who document damp before and after their work.
• Team supervisors who want a consistent standard for their technicians' images.

It is not a tool for homeowners, because a thermal image needs someone who understands emissivity and reflection.

When do you use it?

• During the search: when a patch appears and you want to know its shape and direction before deciding where to listen or which line to pressure-test.
• When documenting: to give the customer a marked-up image and numbers instead of “look at the blue bit here”.
• After the repair: to compare a “before” image with an “after” image from the same angle and with the same range.
• When you suspect a thermal bridge: if the cold sits in corners and where the ceiling meets the wall.

Where do you use it?

• Bathrooms and kitchens: supply routes behind tiles and under floors.
• Ceilings below a bathroom or roof: spreading damp or long linear traces.
• Roofs and facades: provided you shoot away from direct sunlight.

The app runs in the phone's browser at /pro/app/.

What do you need before you start?

• A Kashf Pro account: a free trial once per mobile number, then a monthly or yearly plan or usage credits. Each plan's limits are shown on the plans page.
• A clear thermal image taken with a palette whose name you know: Iron, Rainbow, White hot or Black hot.
• The temperature scale visible in the image, or knowledge of the lowest and highest temperature in the range.
• An optional “before” image if you want to compare after a repair.

How to use it correctly, step by step

Camera settings that make an image fit for analysis

The tool analyses colours, not raw thermal data. The quality of the result therefore starts with how the camera is set up before you press the shutter.

Emissivity and reflection

• Emissivity: most building materials, such as plaster, paint and concrete, sit between 0.90 and 0.95. Set about 0.93 if you do not know the material.
• Shiny surfaces: glazed tiles, polished marble and metal reflect your body heat and the lighting. Shoot at a slight angle, and watch for a patch that moves with you, because that is a reflection, not a leak.
• Reflected temperature: if your camera lets you set it, set it to roughly the room temperature.

Range and palette

• Lock the range manually: turn off auto-ranging and keep the span narrow, for example 5 to 8 degrees around the wall temperature. A wide range hides a small difference.
• The same range before and after: if the range changes between the two images the colours change too, and the comparison becomes misleading.
• Turn off extra layers: visual image blending (MSX), picture-in-picture and coloured isotherm zones all change the colours, and the tool will read them as heat.

Creating a temperature contrast

A leak only shows up when the water temperature differs from the building's.
- Hot line: run the water heater and open the hot water for 10 to 15 minutes, then close it and shoot.
- Cold line: damp cools the surface by evaporation, so it usually looks colder than its surroundings.
- In a Saudi summer, and in other hot climates: water from a roof tank can arrive warmer than an air-conditioned room, so a cold-line leak shows up as warm. Try both modes.
- Facades and roofs: shoot them at night or at dawn, once the solar heat stored in the wall has faded.

A worked example: a living-room ceiling below a bathroom

You photograph a living-room ceiling below a bathroom with the Iron palette, with the scale running from 26.0 to 32.0 °C.

1. You place the rectangle on the ceiling only, choose “Cold spots” and enter 26 and 32.
2. The result: a linear pattern, elongation 4.6, direction 12°, area 6.5% and a difference of 2.1 °C.
3. The route heads towards the washbasin in the bathroom above. You run the hot water for 10 minutes and shoot again in “Hot spots” mode, and a warm line appears on the same route.
4. The preliminary conclusion: a leaking hot line in the bathroom floor. Confirm with a pressure test on the hot line alone, then listen along the route.

After the repair and drying, you shoot from the same angle with the same range and upload the first image as “before”, and the AI writes “✅ Improved after repair”.

Common mistakes that spoil the result

• Leaving the scale bar inside the rectangle: its colours represent the whole range, so it is read as a large anomaly.
• Choosing a palette other than the image's: colour matching fails and the numbers become meaningless.
• Auto-ranging: the camera widens the range when a lamp or window is in view, and the damp contrast disappears.
• An image with visual blending or text over the wall: edges and lines are read as temperature differences.
• Treating cold corners as a leak: corners, ceiling-wall junctions and concrete beams are naturally colder. Compare with a similar dry corner.
• Shooting under an air-conditioning vent: the air stream paints cold patches.
• Comparing before and after before the area has dried: residual damp can last for days after a repair. Wait, or measure with a moisture meter.

The honest limits of the tool

• Photos and screenshots only: the live stream from most thermal cameras is locked inside their own apps, so the tool works on the image after it has been captured.
• It analyses colours, not raw thermal data: converting to degrees assumes the scale is linear between the minimum and maximum. Some manufacturers' palettes differ slightly from the built-in ones, so the difference in degrees is approximate.
• One area: only the largest anomalous area in the rectangle is analysed, so analyse each anomaly with its own rectangle.
• The AI opinion needs a provider connected to the server. If none is connected, the message “AI analysis requires an AI provider to be connected” appears and the on-device analysis keeps working.
• Three AI requests per session: after that, start a new session, which counts against your plan.
• A thermal camera does not see behind the wall: it sees surface temperature only. A deep or insulated pipe may leave no trace.
• The result supports your decision and does not replace it. The shared report itself states that it does not replace a field inspection and professional judgement.

When do you need to escalate?

You are the technician here, but some cases deserve escalation:

• A second opinion from Kashf: when the thermal image contradicts listening and the pressure test, or when the customer refuses to open up without further confirmation. The Kashf team, which works in Saudi Arabia, inspects electronically without breaking, using thermal cameras, acoustic sensing, pressure testing and safe tracer gas, and hands over a documented technical report. Book from the order page or on WhatsApp at +966544027180.
• A licensed structural engineer: if you see cracks in the ceiling, exposed rusting rebar or swollen concrete around the damp area. These are safety decisions that a thermal image cannot settle. See leak damage to concrete.
• An electrician first: if the damp is near sockets, an electrical panel or ceiling lights, switch off the power before any inspection.

Frequently asked questions

Related guides