Abstract
Fruit counting is a fundamental component for yield estimation applications. Most of the existing approaches address this problem by relying on fruit models (i.e., by using object detectors) or by explicitly learning to count. Despite the impressive results achieved by these approaches, all of them need strong supervision information during the training phase. In agricultural applications, manual labeling may require a huge effort or, in some cases, it could be impossible to acquire fine-grained ground truth labels. In this letter, we tackle this problem by proposing a weakly supervised framework that learns to count fruits without the need for task-specific supervision labels. In particular, we devise a novel convolutional neural network architecture that requires only a simple image level binary classifier to detect whether the image contains instances of the fruits or not and combines this information with image spatial consistency constraints. The result is an architecture that learns to count without task-specific labels (e.g., object bounding boxes or the multiplicity of fruit instances in the image). The experiments on three different varieties of fruits (i.e., olives, almonds, and apples) show that our approach reaches performances that are comparable with SotA approaches based on the supervised paradigm.
Highlights
A MONG the multitude of agricultural processes that draw the attention of computer science and robotics researchers, an important role is certainly played by yield estimation
The information encoded by c(i) is clearly ”weaker” than the precise instance count y(i), but it is easier to collect and less prone to human counting errors. Since these labels cannot be used naively to train a counting network, we introduce an image level binary classifier, which will be referred to as presence-absence classifier (PAC) (Presence-Absence Classifier), and use it to train the actual counting network
We proposed a novel weakly-supervised framework for fruit counting in agricultural applications
Summary
A MONG the multitude of agricultural processes that draw the attention of computer science and robotics researchers, an important role is certainly played by yield estimation. While it is possible to reduce the cost and density of yield sampling with the use of autonomous robots to collect images of the orchard [3], [4], [5], the actual fruit counting still remains a challenging task. This is mainly related to the inherent difficulty of extracting high level concepts (i.e., fruits) from raw images due to image anomalies, background clutter, scale variations and occlusions, to name a few. While the detection of some crop varieties is easier due to their shapes or colors (e.g., tomatoes or apples) [6], [7], there are fruit species that can be very difficult to distinguish from background or foliage, such as olives or almonds
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
