STM32n6 bee counting & hornets monitorring system
ABCD - Ai Bee Counting Discrimination system: A bee counting & predation hornets monitorring system using STM32n6-Ai
CALIBRATE IN SEVERAL STEPS:
1st Level:
Video monitoring capable of:
- Detect the opening of the hive or passage to the next ones, or possibly its down falling or overturning!
- Count entries AND exits of bees: for a regular assessment at the end of the day, for example
- Hornet detection at the entrance or around it.
2sd level:
- To complete another method: build another type of audio detection to then allow cross-checking through listening to the buzzing of bees/hornets
DISCRIMINATIONS:
1st level:
- Recognition of individuals by images: bees, hornet, rodent, other
- Scenario discrimination: true/false behaviors through the back and forth movements of bees as opposed to the hovering flight of a hornet.
- Cross-checking with the listening to the buzzing sound and therefore the probable detection of a hornet and constructing another cross-checking criterion.
- Also included if the colony suddenly decreases (collapse due to repeated hornet attacks or illness, for example) : no more “aerial trafic”!
2sd level :
- Unexpected hive opening or overturning (image checking)
- Reduction or absence of bees through the counting in front of the hive entrance
- Hornet rate threshold > absolute detection threshold
- Video detection criteria + audio criteria to consolidate detection and minimize false detections and non-detections
- Backup record detection and/or send alert
1st Level:
Video monitoring capable of:
- Detect the opening of the hive or passage to the next ones, or possibly its down falling or overturning!
- Count entries AND exits of bees: for a regular assessment at the end of the day, for example
- Hornet detection at the entrance or around it.
2sd level:
- To complete another method: build another type of audio detection to then allow cross-checking through listening to the buzzing of bees/hornets
DISCRIMINATIONS:
1st level:
- Recognition of individuals by images: bees, hornet, rodent, other
- Scenario discrimination: true/false behaviors through the back and forth movements of bees as opposed to the hovering flight of a hornet.
- Cross-checking with the listening to the buzzing sound and therefore the probable detection of a hornet and constructing another cross-checking criterion.
- Also included if the colony suddenly decreases (collapse due to repeated hornet attacks or illness, for example) : no more “aerial trafic”!
2sd level :
- Unexpected hive opening or overturning (image checking)
- Reduction or absence of bees through the counting in front of the hive entrance
- Hornet rate threshold > absolute detection threshold
- Video detection criteria + audio criteria to consolidate detection and minimize false detections and non-detections
- Backup record detection and/or send alert

Discussion (0 commentaire(s))