The safe in-field operation of autonomous agricultural vehicles requires detecting all objects that pose a risk of collision. Current vision-based algorithms for object detection and classification are unable to detect unknown classes of objects. In this paper, the problem is posed as anomaly detection instead, where convolutional autoencoders are applied to identify any objects deviating from the normal pattern. Training an autoencoder network to reconstruct normal patterns in agricultural fields makes it possible to detect unknown objects by high reconstruction error. Basic autoencoder (AE), vector-quantized variational autoencoder (VQ-VAE), denoising autoencoder (DAE) and semisupervised autoencoder (SSAE) with a max-margin-inspired loss function are investigated and compared with a baseline object detector based on YOLOv5. Results indicate that SSAE with an area under the curve for precision/recall (PR AUC) of 0.9353 outperforms other autoencoder models and is comparable to an object detector with a PR AUC of 0.9794. Qualitative results show that SSAE is capable of detecting unknown objects, whereas the object detector is unable to do so and fails to identify known classes of objects in specific cases.

https://www.mdpi.com/1424-8220/22/10/3608/pdf?version=1652168325

Anomaly Detection for Agricultural Vehicles Using Autoencoders

Agriculture in Tunisia is a very important economic sector especially olive tree cultivation. Monitoring the latter from space using remote sensing sensors remains challenging and requires to develop adapted code. Detection olive tree size can help to monitor the growth. To do it, we propose in this work to detect tree crown on high resolution satellite images, in particular Pl&#x00E9;iades. To achieve this goal, we proposed the use of the Deep Learning approach Detectron2 in order to produce instance image segmentation, to make distinguishing between the olive tree, its shadow and the soil. Detectron2 was trained using a database of realistic images generated by the DART model. The evaluation results prove the validity of our proposed approach also the visual inspection shows good agreement with the reality. 

Olive tree and shadow instance segmentation based on Detectron2

The Earth Observation-based Anomaly Detection (EOAD) system: A simple, scalable approach to mapping in-field and farm-scale anomalies using widely available satellite imagery

Yeison Rodriguez

Papers

The Earth Observation-based Anomaly Detection (EOAD) system: A simple, scalable approach to mapping in-field and farm-scale anomalies using widely available satellite imagery