Colour sorting is a vital process in manufacturing of high-quality wood products, however hitherto a relatively manual process in majority of facilities in Malaysia. Automation is an ideal solution; however, costs are prohibitive for small-medium industries (SMI). Thus, this project aims to produce a flexible solution that can cater for different scales of manufacturers. Three cameras of different price ranges are used: i) Hikrobot® MV-CE200-10UC (Hikrobot), ii) Logitech® C920 (Logitech), and iii) Sony® RX0 II (Sony). Having set up a veneer imaging prototype, human sorted images of American red oak (Quercus rubra), yellow poplar (Liriodendron tulipifera), and maple (Acer spp.) are acquired. After performing image preparations and calibrations, 26 features are extracted from each image. The features are based on the average and standard deviation of the wood basal colour and wood grain colour. The salient features are obtained using the Sequential Forward Selection (SFS) method. The most salient features are then used to train a Self-Organizing Map (SOM), and the resulting maps are observed. From results, it is affirmed that the colour of the basal colour is highly correlated with human sorted colour groups. As expected, Hikrobot topped the charts being of industrial grade. It is interesting that Logitech exhibited comparable performance. Sony performed the worst due to software limitations. This proposed system can achieve accuracy of 89.0% for red oak, 94.3% for poplar and 96.4% for maple. This research will assist the SMI to develop an affordable vision system for colour sorting.