Products

FLAGSHIP

K-스마트 수목관리 솔루션

IoT 센서와 AI 진단 모델로 도시 가로수의 생육 상태와 병해충을 예측하는 스마트 수목관리 플랫폼. 모듈식 가로수 보호 장치를 지자체 현장에 설치하며, 사후 대응형 관리를 데이터 기반 예측 관리로 전환하고 있습니다.

IoTLPWAN/USIMAI DiagnosticsFaster R-CNNYOLOXDigital Twin

3D 관제 대시보드

개별 수목 3D 시각화·활력도 정밀 분석·병해충 예측 알림을 제공하는 관제 시스템입니다.

Species-relative assessment — group baselines instead of absolute thresholds

At the same 20% soil moisture, a drought-tolerant pine group reads as normal while a cherry-tree group reads as water-stressed. Diagnosis is made from the deviation (magnitude and duration) against the reference growth model of a tree profile group defined by species, region and soil environment. Applying group-relative assessment instead of a uniform threshold comparison prevents species-specific misdiagnosis.

SW · HW role split

Software — AI platform

Lumir Inc.

We build AI diagnosis of tree growth data, the 3D control dashboard and citizen-participation services.

Hardware partner

Treemate Inc. (hardware partner)

A partner that manufactures and installs street-tree protection equipment. Procurement and installation records belong to Treemate.

From analog management to digital transformation

AS-IS
Reactive response — action only begins after dieback or fall has occurred.
TO-BE
Data-driven prediction — AI catches anomalies first, instead of visual inspection.
AS-IS
Labor-dependent management — there are not enough people to survey trees exhaustively.
TO-BE
Citizen-participation care — citizen reports feed directly into the management system.
AS-IS
Climate risk — pest surges and soil hardening depress growth.
TO-BE
Preemptive management — early diagnosis and prevention before problems grow.

Actual product screens

The screens below are the real product UI; the tree and alert figures shown are demo simulation data.

Control dashboard — a map view of managed trees per municipality with risk grades (high/medium/low) — Control dashboard

Species-specific comparison — at the same 22.6% soil moisture a pine reads normal while a cherry tree reads water-stressed — Species-relative assessment

Alert center — absolute-threshold alerts shown separately from group-relative alerts — Alert center

Tree detail — per-tree diagnostic evidence and deviation trend against the reference growth band — Per-tree diagnosis

Control center — coverage view separating alerts that can be handled automatically from those needing site inspection — Remote control coverage

Citizen report screen — reached by QR scan, with a form that collects no personal data such as name or contact — Citizen reporting

Deployment status

Hardware installed by our partner Treemate Inc. is in place at 15 sites across 11 municipalities. Most are landscape trees in parks and public buildings rather than street trees, and the public-procurement registration and installation record belong to Treemate. Lumir is building the AI diagnosis platform on the data collected from these sites.

Target metrics

AI diagnostic accuracy 90%+ (target)
Pest damage detection rate 80% (target)
Early detection of growth anomalies 80% (target)
Management time -20% (target)

Considering adoption?

We will walk you through how it applies to municipalities, public institutions and landscape management sites.