Two prototype automatic grading systems for dried shiitakes are presented. For the first prototype, image processing algorithms utilizing neural network were developed and implemented. Image processing utilized the raw image of fed dried shiitakes without any complex processing such as feature enhancement and extraction. Performance of the network based grading was presented. Another prototype has been developed considering the efficiency of the practical usage of the system. Quantitative visual descriptors were used for grading. Grading criteria was controlled interactively. Both prototypes were composed of automatic devices for mushroom feeding and handling, a set of computer vision system with lighting chamber, one chip microprocessor based controller, and pneumatic actuators. The network based software and related interface have been developed, which can remotely control and manage an on-site operating system. Developed software modules were composed of two parts: monitoring/management modules and control/diagnosis modules. And they were designed to run on the internet and local network.