Calorific value estimation method, calorific value estimation device, and waste storage facility

Abstract

A captured image of waste in a pit is divided into a plurality of prescribed cells, a brightness histogram of each cell is prepared, the cells are classified based on the brightness histogram into a plurality of labels using prescribed classification criteria, and to a certain label from among the plurality of labels, an evaluation value is assigned representing a calorific value to be generated when waste in the cell classified to said label is incinerated in an incinerator. A group of labels to which the evaluation values have been assigned up to this point by learning is clustered into a prescribed number of clusters based on the evaluation values, a weight representing an incidence of each label in numerical form using an optional correction value is obtained for each of the clusters from the results of the clustering, and a cumulative value obtained by normalizing the weights obtained by repeating the aforementioned process is set as an expected value. The expected value of any label is obtained for each cluster, and the calorific value of waste in a cell classified to said any label is estimated based on the evaluation value for the cluster with the highest expected value for said any label.