This letter presents an incoherent change detectionalgorithm (CDA) for wavelength-resolution synthetic apertureradar (SAR) based on convolutional neural networks (CNNs).The proposed CDA includes a segmentation CNN, whichlocalizes potential changes, and a classification CNN, whichfurther analyzes these candidates to classify them as real changesor false alarms. Compared to state-of-the-art solutions on theCARABAS-II data set, the proposed CDA shows a significantimprovement in performance, achieving, in a particular setting,a detection probability of 99% at a false alarm rate of0.0833/km2