Multispectral imaging using LED illumination and an RGB camera

LED illumination based multispectral imaging is getting much attention in recent years due to its fast computer controlled switching ability, availability of many different LEDs, robustness, and cost effectiveness. In this paper, we propose a system which uses an RGB camera along with two or three combinations of three different types of LEDs in order to acquire multispectral images of six or nine channels. Optimal LED combinations are selected so as to produce accurate estimate of spectral reflectance and/or color. The system is rather simple to realize. Moreover, it is faster as it requires only two or three shots, unlike state of the art multiplexed LED illumination based systems which require as many shots as the number of channels that a system can acquire. The proposed system can be useful in general multispectral imaging applications. The system has been evaluated with both the natural images and paintings. The results from the simulation experiments were promising, indicating the possibility of the proposed system as a practical and feasible method of multispectral imaging. Introduction Multispectral imaging has many advantages over traditional three channel (usually RGB) color imaging. It is less prone to metamerism, produces higher color accuracy, and unlike digital cameras, it is not limited to the visual range, rather can also be used in near infrared, infrared and ultraviolet spectrum as well, depending on the sensor responsivity range. Spectral reflectance of a scene which represents the unique property of an object, can be recovered from the images acquired with the spectral imaging systems. Multispectral imaging, therefore, has widespread application domains, such as remote sensing [1], astronomy [2], medical imaging [3], biometrics [4], culture and heritage [5, 6] and many others. Many different types of multispectral imaging techniques and systems have been proposed in the literature. In a typical filter-based imaging system, either a set of traditional optical filters in a filter wheel, or a tunable filter [7, 8], or in front of a high quality digital camera [9–11] or a stereo camera [12, 13] are employed. Multispectral filter array (MSFA) [14–17] is another type of multispectral imaging technique which extends the filter array from 3-channel like in Bayer pattern further, allowing to capture more than 3 bands. Another promising technique of multispectral imaging, which is of our primary interest in this paper, is based on a multiplexed LED (Light Emitting Diode) illumination [18–20]. In a typical LED illumination based multispectral imaging system, a set of n different types of LEDs are selected, each type of LED is illuminated in a sequence, and a monochrome camera captures an image under the illuminated LED, thus producing a n band multispectral image. Such a system modulates the illumination and provides a multispectral light source. LED illumination based multispectral imaging has been used in several applications like biometrics [4], medical imaging [3] and film scanner [21]. This has got much attention in recent years because of the advantages of the LEDs: fast computer controlled switching ability, robustness, and cost effectiveness. Availability of many different color and high intensity LEDs with peak weavelengths spanning the whole visual range and even infrared region has made the construction of more effective multispectral system possible. Shrestha et al. [21] recently proposed a LED based spectral film scanner. Tominaga and Horiuchi [20] proposed multispectral imaging by synchronizing capture and illumination using programmable LED light sources. Studies are being done to anayze and address different issues with the LED based multispectral imaging. Spectral variability of LEDs with angle and time of usage is one example of such studies, done by Martinez at al. [22]. They found that white light LEDs designed with a blue emitting LED coated with a yellow emitting phosphor, emit light whose spectrum changes as a function of the angle and time. On the other hand most of the single color LEDs are found to be invariant compared to the white lights, and hence they are recommended to be used in LED based multispectral systems. Shrestha and Hardeberg [23] has proposed a binary tree based LED matrix/panel design method which produces an optimal or suboptimal arrangement of LEDs for equal energy and uniform ligthing in a LED based multispectral imaging system. Most of the state of the art LED based multispectral imaging systems require many different LEDs and many shots (n shots for n types of LED) in order to acquire a multispectral image of a scene, and this also makes the systems complex, limiting their practicability and feasibility. In this paper, we propose an efficient LED illumination based multispectral imaging system that requires fewer shots, using an RGB camera. Instead of lighting a LED at a time sequentially, combinations of three optimal LEDs are selected, each combination of LEDs is lit at a time, and a three band image is captured by the camera each time. With n shots, we can acquire 3n band multispectral images. The system is practical, and easily and cheaply realizable. It can be used as a general multispectral imaging system in many different applications including in culture and heritage artworks. We present the results with the natural images, and painting images using a famous painting, the Scream, by Edvard Munch from 1893 as an example. After this section, we present next the proposed system and the methodology. We then present experiments and results. The results will be discussed next, and finally we conclude the paper. 8 © 2013 Society for Imaging Science and Technology The system and methodology The proposed LED illumination based multispectral imaging system comprises of an RGB camera, LED panel(s) and a microprocessor based controller (or a computer), as illustrated in Figure 1. A LED panel is built with combinations of three different types of LEDs, or alternatively, more than one LED panel, with each panel comprise of three types of LEDs can be used. We assume that the LED panel is made in such a way that the resulting illuminat ion from the panel is uniform.