Smart Cameras
In addition to its standard GigE camera family, GEViCAM also specializes in incorporating hardware neural networks in GigE Vision cameras. The combination provides stand alone intelligent sensors which can communicate via a high speed network. GEViCAM smart cameras are completely different from traditional CPU controlled or embedded DSP smart cameras. They learn from experience just like a human brain and do not employ a CPU or micro controllers inside the camera. The GigE platform is essential for smart cameras even though for most of the time they function in a stand-alone mode. However, once a smart camera detects a reporting event such as a defect, inspection error, surveillance incidents, etc., it then must keep sending multiple images without missing a frame. In this case, video data band-width must be high and GigE is the best platform for those applications. GEViCAM smart cameras (GN-Series) are designed in the same package as its standard GigE cameras but nevertheless contain the neural network brain within the same small envelope.
|
GN-Series (with Neural Network Chip)
|
[GN-Series]
| GN-3000 (1/3", WVGA, 60 fps., B/W) - Preliminary
GN-3000C (1/3", WVGA, 60 fps., Color) -Preliminary
|
Neural Network Camera
Prior to the release of GEViCAM's own Smart Camera with Gigabit Ethernet Technology, we introduced that the MTVS (V1KU) "industrial-strength" camera and optional robust enclosure. The MTVS was designed and developed by General Vision and is available now.
| |
| The MTVS is the smallest, fastest, and most affordable trainable vision device available. This is a production-efficient solution that allows savings on the inspection costs in many ways.
M: Miniature
T: Trainable
V: Vision
S: Sensor
Features:
- Lower-power consumption
- No PC needed
- Higher throughput and higher accuracy than human inspection
- Flexibility for varied lighting environments
- On-site fine adjustment
- Easy to capture image
- Easy to train
- Easy to run
- Easy to collect results
- Easy to retrain or augment training
|
Sensor: 1/3" CMOS, Progressive Scan, B/W
Resolution: 752x480
Frame Rate: 60 fps.
Mechanical: 30 x 30 x 60 (mm), 50 g
Neural Network Chip: Cognimem CM1K(1024 neurons)
Communicatoin: High Speed USB
|
| Applications:
- Part Inspection
- Model Identification
- Classification
- Quality Check
- Defect Detection
- Label Check
- Object Tracking
- Face Recognition
- Conformity Verification
- Condition Monitoring
- Event Detection
- Surface Inspection
- Alignment/Positioning
| |
|
