![]() In some cases, these environments could be similar in terms of lighting conditions. Therefore, it is necessary to capture the image of the color reference target and the target object in two different environments for training and testing purposes. A reliable model would need as many samples in the sRGB color space as possible. However, most color reference targets were not waterproof. The best practice in developing a color conversion model is to obtain training and testing data in the same environment for image capture. As a result, fish health might be affected, and fish loss might occur. It might also be anesthetized for easy color measurement and subjected to a few color measurement sessions during the culture period. Remarkably, a fish should be moved out from the aquarium to a dry plate for measuring its color using a colorimeter. Although an evaluation panel could be established to validate the results of the calibrated model, the panel might find it challenging to deal with an uneven color distribution on the surface area of small ornamental fish. When assessing the model accuracy, the colors measured by the model were generally compared with those from a colorimeter, regardless of the disadvantages experienced with the translucent sample matrix or any uneven color distribution. In a CVS, an RGB to L*a*b* color conversion model was trained by fitting the RGB color values of some color patches with their known L*a*b* colors. The development of a CVS is straightforward and has been described in various reports. Due to the high spatial resolution, a CVS would also facilitate color distribution analysis. Therefore, vision-based color measurement is critical for small or colorful ornamental fish such as clownfish or rainbow fish. The pixel size can be about 0.5 mm or even less depending on the image resolution, but it is relatively much smaller than the 8-mm aperture diameter of a popular Minolta colorimeter. A similar phenomenon might happen with a thin layer of skin whose measured color would be significantly affected by the meat matrix underneath the skin layer.Īnother significant advantage of a CVS over a colorimeter is that it can measure a sample surface as small as the size of a pixel. The possible reason was that the CVS allowed lower light penetration resulting in less refraction or optical discontinuities. The CVS measurement was more accurate than a colorimeter in measuring a meat sample with a translucent matrix. ![]() A CVS would also be a promising tool for rapid analysis of color distribution.Īs a rapid, consistent, and non-invasive measurement approach, the CVS is gaining more popularity in color-related research. This favorable feature showed great potential for monitoring dynamic processes in real-time, such as thermochromic sample analysis or roast rice color characterization. Ī CVS could perform a non-contact color measurement in real-time. A more innovative approach was to develop a computer vision system (CVS), for which a regression model was trained with the image of a color reference target to measure the fish color mainly in the L*a*b* color space. The colorimetric analysis could also be simply performed based on image analysis to detect the color changes with respect to a standard color value. Conventionally, a colorimeter was used to measure the fish's color after specific periods of special feeding,. Therefore, special nutritional supplements were applied to fish farming to control or enhance fish pigmentation, ,, ,,. Like other food products, the color of a fish or fish product is a decisive visual attribute affecting the customer’s purchase. Preliminary results of measuring the clownfish color showed that the proposed system had much potential for in situ and non-invasive visualizing of the color distribution of small ornamental fish with color non-uniformity. Consequently, an optimal cubic ridge model was obtained with appreciably low training and testing mean color errors of 1.17 ± 0.64 and 1.10 ± 0.50 CIEDE2000 color difference units, respectively. The generalization ability of the model was ensured with regularized regression. Brightness adjustment was effectively applied to these images to minimize uneven illumination in both environments. The research goal was achieved by developing a color conversion model trained with images of a color reference target in the ambient environment and clay samples in the aquarium representing the target fish. ![]() This study proposed an innovative technique to measure the fish color in situ based on machine vision to avoid potential effects on fish health that might result from conventional measurement with a colorimeter.
0 Comments
Leave a Reply. |