System load
Track CPU utilisation and used versus total RAM once per second.
// runtime evidence · five-minute rolling history
Embedded Vision Performance Monitoring
Compare FPS, CPU, memory and thermal behavior while camera pipelines run on their target device.
A pipeline that produces the right image can still miss its target because of CPU load, memory pressure, thermal behavior or insufficient frame rate.
1 Hz sampling5-minute historyPer-camera FPSCPU, RAM and thermals
Thermal state
Observe SoC and CPU temperature values exposed by the target system.
Camera throughput
Monitor per-camera FPS, running state, resolution, frame size and frame count.
Network activity
View transmitted and received throughput alongside camera and system measurements.
Rolling history
Keep up to five minutes of one-second samples for immediate visual comparison.
Pipeline timings
Inspect execution time and previews for individual CV nodes while iterating.
// workflow
From hardware to defensible evidence.
- 01
Establish a baseline
Run the intended camera and processing path on the target hardware.
- 02
Change one variable
Adjust pipeline configuration, processing parameters or camera settings.
- 03
Compare the evidence
Review FPS, load, memory, temperatures, network activity and node timings.
// technical evaluation
START INQUIRY →Review your target and sensor combination.
Share the board, Linux image, sensor and current engineering objective. Pixoric will scope the evaluation setup with you.