Trial Project |
iCremation
(REF: P-0140) |
Matched I&T Wish |
Development of AI-based Image Analytic System for Cremation System (REF: W-0219) |
Matched I&T Solution |
Cremation Process Automation System (REF: S-0236) |
Solution Feature |
- Cremation Automation System analyses real time cremation images and automatically control air-flaps to fully automate the cremation PLC control process. Optimization goal is to reduce the overall cremation time and pollution/emission.
- The system consists of: 1. Input IP cameras, which capture input images through live streaming. 2. Computation GPU Server, which runs a Convolution Neural Network that will classify the camera images to determine the state of the cremation process and responses. 3. Air-flap Control Output GUI to control PLC.
- The first stage of the project will involve data collections and CNN model design and development. Image and coffin content data will be collected from all 6 chambers for training (to avoid over-fitting to certain chamber’s data set)
- The second stage will involve model training and trial runs. Initially trained model will be used to provide AI generated air-flap control outputs and be imported into PLC for automated controls. Data will be collected through trial runs and results will be collected for further optimization.
- The final stage involves analyzing the initial model’s results and collecting additional training samples for further optimization on cremation time and emission. We will correlate the results with coffin/body weights, type, sizes, preservative level to identify trends and optimization strategy.
|
Trial Information |
Trial Site |
: |
Wo Hop Shek Crematorium |
Trial Duration |
: |
Feb 2020 to Feb 2021 |
|
Final Report |
EMSD-M&V-P0140-W0219-S0236_Final.pdf
|
Info on I&T Solution Provider |
Solution Provider | : | Hampen Technology Corporation Limited | Address | : | 5B, Mackenny Centre, 660 Castle Peak Road, Cheung Sha Wan, Kowloon, HK | Contact Person | : | Felix Chow |
Position | : | CEO | Tel | : | 9186 3684 | Email | : |
fchow@hampentech.com | Webpage | : | www.hampentech.com |
|