If you already develop a farm management platform or a vision model, access to original insect images can matter more than a second AI dashboard. A useful pilot connects the trap, camera and receiving system and evaluates the images before promising automated counts.
Choose the trapping method first
A sticky-board camera project and a light-trap monitoring instrument collect different samples. Confirm the target insects, crop, season and whether yellow or blue boards are needed. Specify board dimensions, replacement access and the camera-to-board geometry. A product photograph does not prove that a particular board, focus distance or target species is supported.
Ask for original sample images
Evaluate actual JPEG files at the intended distance and lighting, rather than compressed screenshots. Check whether small insects are distinguishable, whether all parts of the board are in focus and whether glare or dust obscures the surface. Record camera resolution, capture schedule and board age with the sample. More pixels alone do not establish recognition performance.
Separate capture from inference
Decide whether your platform receives original images, counts, or both. Agree device identity, capture timestamp, board reference and optional battery and signal fields. For customer-owned AI, direct image delivery must be scoped with the receiving endpoint and verified on the selected firmware. Do not treat RTSP video, scheduled JPEG upload and a cloud-result API as the same capability.
Validate counts for the deployment
Use representative images from the intended crop and site and have people label the target classes. Report false detections, missed detections and species confusions, including images with overlapping or damaged insects. Keep an unreviewed model output distinct from a confirmed result. A detection count is not automatically a treatment threshold; local agronomic review remains part of the workflow.
Measure power and field service
The complete load includes capture, illumination if used, image processing, cellular wake-up and retransmission. Specify the worst reporting day when sizing a solar system. Confirm component temperature ratings and whether a battery can charge at the minimum temperature. Enclosure ratings, board replacement, cleaning and condensation should be evaluated on the actual assembled design.
How to start with us
Our catalogue includes several light-trap instrument configurations. A compact sticky-board camera and direct delivery to your own AI workflow should be treated as a project with a confirmed mechanical design, data contract and quotation. North American radio bands, FCC/ISED documentation and extreme-cold performance must be established for the supplied configuration. We do not present a development proposal as an already-certified deployment.
Before requesting a quotation
Send a clear project brief.
- Target insects, crop, board colour and board dimensions
- Required sample JPEGs, capture interval and upload method
- Your own AI, optional supplier analysis, or images only
- Country, component temperature limits, power and service interval
Continue your project
Explore pest monitoring equipment and project options ↗
Explore optional ZhiCloud AI workflows ↗
Ask Icey about your configuration ↗