Abstract
For machine-vision systems, the image quality is the major factor directly affects the final image processing results. Especially under the natural lighting conditions, image quality varies significantly as the light source condition changes. This light source condition changing is inevitable in crop field remote sensing applications. In this paper, an artificial intelligent algorithm was developed to realize automatic adjustment of the camera settings, such as the "gain" and "exposure time" to compensate the unstable ambient illumination for improving image quality. Therefore, we could detect as much information of the target as we desire. Using a DuncanTech MS4100 highresolution digital color and multispectral camera as the vision sensor, we implemented the real-time white balance of the image among three channels (blue/green, red, near-infrared) under different natural lighting conditions.
| Original language | English (US) |
|---|---|
| State | Published - 2005 |
| Event | 2005 ASAE Annual International Meeting - Tampa, FL, United States Duration: Jul 17 2005 → Jul 20 2005 |
Other
| Other | 2005 ASAE Annual International Meeting |
|---|---|
| Country/Territory | United States |
| City | Tampa, FL |
| Period | 7/17/05 → 7/20/05 |
Keywords
- Adaptive neuro-fuzzy inference system
- Artificial intelligent
- Artificial neural network
- Auto white balance
- Machine vision
- Remote sensing
ASJC Scopus subject areas
- General Agricultural and Biological Sciences
- Bioengineering