Context-Based Visual-Language Place Recognition