An industrial camera for color image processing with mixed lighting conditions

Industrial camera is an important electric device in the machine vision, which serves as an interface between the computers and the real world. There is an increasing demand on the usage of high-speed, miniaturization, portable, low-power, and smart industrial camera. In this paper, we novelly design and develop such a kind of camera for capturing and processing color images from the real world industrial environment with mixed lighting conditions. Besides, a color image processing flow has been embedded in the camera to gain high quality images. For the seek of the robustness of the proposed camera to lighting variations, we improve the algorithm of automatic white balance (AWB). The improved algorithm achieves a sound white balance processing result on real world images. Finally, we realize all the processing flows and algorithms in the embedded system and apply it to practical applications.

I. Introduction

In the field of machine vision, there has been a growing interest in CMOS image sensors. The most probable reason for this growing interest is the demand for miniaturization, portable, low-power, and cost-effective imaging systems. One of the most advantages for utilizing CMOS may be concluded as it is able to be integrated. A single CMOS-on-a-chip can be integrated with a number of electronic chips as analog-to-digital converter (ADC) and full digital interface for some specific purpose [1]. As an industrial application, there are trends for the combination of CMOS image sensor and DSP high-speed chip to take images under complicated industrial environment [2]. This new electronic device can be applied in many fields, such as industrial pipeline, Intelligent Transportation Systems (ITS), video monitor, etc. The high-speed computational property of DSP ensure us to take some image processing algorithms which were ever implemented on computers into the embedded system for real time processing [3].